Introduction

This paper presents experimental validation of a technique for monitoring pollution plume geometry during physical modelling in a laboratory environment. The method is non-invasive and, therefore, does not disturb the soil fabric or distort the evolving contaminant flow. Monitoring is achieved by constructing a standard Wenner array of electrodes at a small scale. This array configuration is used to produce two-dimensional (2-D) images based on resistivity tomography. The main advantage of this method is that the miniaturised probes are inserted into the base and on the surface of the physical model, so that the instrumentation does not compromise the experiment.

In order to develop this physical model in the laboratory floor, several field data were used, as they derived by an in-situ campaign at an old domestic waste tip with no lining system and no environmental monitoring for methane gas and leachate production. This waste tip is situated in the city of Patras, Hellas and acts as a dilute and disperse site in which attenuation of leachate within the body of waste and in the unsaturated zone of the underlying strata ameliorates the impact of groundwater pollution. A partially saturated zone of alluvial fan deposits underlies the site, and this retards leachate percolation to the underlying water table. The site conditions revealed during the geological and geophysical exploration provided the basic components for the development of the generic model described later in this paper.

The miniaturised electrical imaging technique developed for imaging contaminant plumes in generic models represents a direct analogy to field-scale electrical imaging surveys similar to the one undertaken during the geophysical exploration of the Patras waste tip. The main objective of this paper is to demonstrate the miniaturised electrical imaging method and discuss its possible applications in problems associated with physical studies of contaminant plume evolution in a laboratory environment.

Site description

The discussed landfill site served the waste disposal demands of the city of Patras during the period 1971–1993. It occupies an area of 237,000 m2 and it is estimated that 800,000 ton of waste have been deposited during its operation, with a mean compaction of 0.5 ton/m3 (Depountis et al. 1998b). The composition of the deposited wastes is mainly organics, paper, plastics, rubble, and rubbers. At present, the site has been restored according to the European environmental standards. Prior to the site restoration, a temporarily unsuitable capping material was used to prevent water infiltration into the waste tip. The inadequate permeability of the capping material was the main reason for leachate generation and migration beneath the landfill base.

Geological conditions

The substructure, upon which the refuse has been deposited, consists of river bed material and alluvial fans. According to Depountis et al. (1998a), the geological formations in the area can be classified with the following sequence from the youngest to the oldest (Fig. 1).

  1. 1.

    Landfill site: modern formation consisting of domestic wastes with the inactive component being mainly rubble (6%). The active component is mainly organics (54%), paper (22%), plastic, and rubber (10%).

  2. 2.

    River bed deposits: these are recent loose coarse-grained deposits with a thickness of 1–2 m, consisting of boulders, cobbles and gravels, and a small percentage of sand.

  3. 3.

    Alluvial fan: the fan deposits are mixed facies of cobbles, gravels, pebbles, and fine-grained material consisting of clayey sand. According to Koukis and Depountis (1995), the percentage of the fine-grained material is 40%. These sediments are overlain in some places by terraces with the same lithological characteristics. The fan-terraces deposits have a thickness up to 100 m. This formation is highly permeable with a hydraulic conductivity K=10−5 m/s to 10−2  m/s. The conductivity decreases at deeper horizons as they become more compacted.

  4. 4.

    Alluvial accretions: at the lower part of the investigation site alluvial accretions overlie the basal Plio-Pleistocene bedrock. They consist of clayey silts, well-graded sands, and loose to semi-compacted gravels and cobbles. They range up to 200 m in thickness and they are considered to be semi-permeable to permeable (Rozos 1991). This formation constitutes the main aquifer of the area.

  5. 5.

    Weathered mantle: this is a recent formation consisting of weathered Plio–Pleistocene sediments. It comprises of weathered gritty marls with a grey–yellow colour and ranges in thickness from 2 m to 5 m.

  6. 6.

    Diluvial formations: these formations occur outside the investigation area and overlie the bedrock in limited sections. They consist of poorly sorted silty gravels, cobbles, and pebbles with the cobbles and pebbles being loosely cemented by red–brown clay.

  7. 7.

    Plio–Pleistocene sediments: these sediments form the bedrock of the enclosed area of the city of Patras. They are mainly fluviolacustrine, lacustrine, and marine sediments consisting of fine-grained facies of clayey or gritty marls and sparse lenticular intercalations of conglomerates of thickness 0.1–1 m. In many places the conglomerate thickness ranges between 3 m and 5 m. These sediments are considered to be semi-permeable formations.

Figure 2 illustrates a geological cross-section across the landfill site (based on Fig.1) in East–West direction.

Fig. 1
figure 1

Geological map of the landfill site

Full size image
Fig. 2
figure 2

Cross-section of the landfill site

Full size image

Application of electrical imaging under in situ conditions

Electrical imaging (resistivity tomography) was used in order to investigate potential leachate migration from the landfill base. A single multicore cable (Campus imager cable), to which 25 electrodes were connected at fixed takeouts, was linked to a switching module in the earth resistance meter (Geopulse system) that was controlled by a computer. The Geopulse meter was operated using the control software Image25. The survey design was saved in a parameter file, which could be accessed by the Image25 whenever needed, and all the resistance measurements during the survey were stored in memory for later processing. Barker (1997) discusses in detail the principles and applications of this technique in field-scale investigations.

The survey was carried out prior to the site restoration, mainly in locations where the temporary capping material had been contaminated by leachate seepage, which was occurred from perched leachate tables generated between the body of waste and the daily cover soils.

Several measurements were made using arrays of 25 electrodes spaced either at 8-m or at 5-m intervals in a Wenner array configuration. Four different lines of electrodes were used in a close grid with each line having a total length of 192 m (Fig. 3).

Fig. 3
figure 3

Resistivity survey lines

Full size image

Two different lines of electrodes (lines 5 and 6) having a total length of 120 m each were used during a wet period in order to explore any potential difference in leachate generation. Lines 5 and 6 were undertaken at the same positions as lines 3 and 4, respectively, but with a smaller spacing (5 m) in order to obtain a better resolution of the internal structure of the waste tip.

The observed resistance values were used to plot 2-D images of the subsurface by using the RES2DINV computer program. This automatically determines a 2-D resistivity model for the subsurface using the data obtained from electrical imaging surveys (Griffiths and Barker 1993) and the final processed images provide a good representation of the model subsurface resistivity. A forward modelling subroutine is used to calculate the apparent resistivity values, and a non-linear least-squares optimisation technique is used for the inversion routine (Loke and Barker 1995, 1996).

All cross-sections (tomographies) undertaken in the field area show low resistivity zones, which correspond to the presence of leachate (Figs. 4, 5). The scale shown in these figures is an apparent resistivity scale measured in Ωm (ohm.m) Values below 5 Ωm characterise the presence of perched leachate tables (leachate pockets) within the waste tip. Values between 5 Ωm and 20 Ωm characterise geological materials (either refuse or fan deposits), which have been saturated by the leachate percolation. Values above 20 Ωm characterise the partially saturated fan deposits. The depth to the base of the landfill was already known in places from geotechnical boreholes previously drilled at the site. In conjunction with the resistivity tomographies, it was possible to determine with a good accuracy the base of the landfill as can be seen in Figs. 4 and 5 (continuous black line).

Fig. 4
figure 4

Resistivity tomographies of the landfill site (Lines 1, 2, 3, and 4 with an electrode spacing of 8 m, respectively)

Full size image
Fig. 5
figure 5

Resistivity tomographies of the landfill site (Lines 5 and 6 at the same positions as lines 3 and 4, respectively, but with an electrode spacing of 5 m)

Full size image

Detailed interpretation of tomographies

Line 1 (from East/Southeast at 0 m to West/Northwest at 192 m)

The average thickness of the deposited wastes is 11–12 m. Leachate pockets occur at several positions within the body of the wastes. Leachate is present within the landfill and has migrated below the landfill base northwestwards, following the topography of the site and a channel structure.

Line 2 (from Southwest at 0 m to Northeast at 192 m)

The average thickness of the deposited wastes is 8–9 m. Leachate is present along the body of the landfill and has migrated 5 m below the landfill base.

Line 3 (from Southwest at 0 m to Northeast at 192 m)

The average thickness of the deposited waste is 7 m. Leachate does not seem to have percolated below the base of the landfill.

Line 4 (from Southwest at 0 m to Northeast at 192 m)

The average thickness of the deposited wastes is up to 7 m. Leachate is present along the body of the landfill and has migrated below the base of the landfill.

Line 5 (same position as line 3, but with an electrode spacing of 5 m)

The line has a total length of 120 m, and the first and last station corresponds to 40 m and 160 m of line 3, respectively. The smaller spacing reveals that leachate migration has progressed below the landfill base. At the centre of the 2-D tomography model (Fig. 5) the development of a leachate plume, possibly saturating the clayey sand material of the fan deposits can be clearly detected.

Line 6 (same position as line 4, but with an electrode spacing of 5 m)

The line has a total length of 120 m. The first and last station corresponds to 40 m and 160 m of line 4, respectively. Leachate pockets occur within the body of the wastes. Between 95 m and 125 m, on the horizontal scale, to a depth of 14 m below the surface, a large leachate pocket is apparently migrating beneath the landfill base, possibly saturating the fine-grained materials.

Resistivity lines 1–4 were surveyed during a dry period, whereas, resistivity lines 5 and 6 were completed during a wet period.

Summarising, it is evident that the low resistivity anomalies shown in Figs. 4 and 5 represent leachate pockets and plumes, which migrate towards the underlying alluvial fan deposits. This is consistent with the knowledge that the landfill is not sealed. This suggests that the contaminant fluid could migrate longer towards the underlying water table.

Application of electrical imaging under model conditions

Miniaturised electrical imaging equipment

Based on the field-scale resistivity apparatus, described above, a miniaturised electrical imaging equipment was developed for detecting and plotting contaminant plume evolution under model conditions in a laboratory environment (Depountis et al. 1999). The main elements of this apparatus are:

  • Two 25-electrode arrays located at the surface and the base of the model. The two arrays allow two individual resistivity surveys, one from the surface downwards and the second from the base upwards. The electrodes consist of brass pins with a length of 0.8 cm and are connected to individual cores of two screened 25-way cables. The spacing between the pins is 3 cm.

  • The screened cables are connected to a relay box via a connection box situated at the top a centrifuge strongbox in which the physical model is assembled. A 12 V DC power supply unit operates the relay box, activating either the surface or the basal array.

  • A multicore cable connects the relay box with the Geopulse earth resistance meter. The resistance meter can be operated from a PC by using the control software Image25 and the resistance readings are processed with the RES2DINV program. After processing, the data can be obtained in a xyz format (x=length, y=depth and z=apparent resistivity). Thus, by using a contour package (e.g. Surfer) a contoured cross-section of the examined model can be plotted.

Apparatus for contaminant infiltration tests

A generic model of contaminant infiltration into partially saturated sand was designed in order to explore the potential of testing contaminant plume evolution using the miniaturised electrical imaging equipment. Tests were undertaken at the laboratory floor in order to evaluate the equipment’s efficiency for similar bench and centrifuge tests. The apparatus (Fig. 6) used for this kind of tests is the following:

  • Rectangular centrifuge strongbox filled with pluviated sand, thus creating a sand model with dimensions: length=79.5 cm, width=35.5 cm, and height=18 cm

  • Aluminium liner inside the centrifuge strongbox with perforations in the base and sides to enable uniform saturation of the pluviated sand.

  • Miniaturised electrical imaging equipment.

  • Perspex cylinder (0.5-m high) acting as the contaminant reservoir.

  • Solenoid valve to control the contaminant flow between the reservoir and the sand model.

  • Perforated distributor tube to allow uniform distribution of contaminant across the centre line of the model.

  • Two matrix potential probes (tensiometers) placed inside the model to monitor the capillary pressure and therefore, the degree of saturation during the experiment.

Fig. 6
figure 6

Apparatus for contaminant infiltration tests

Full size image

The soil material used for the sand model was Congleton Sand, which is poorly graded with d10=100 μm and a coefficient of permeability K=10−4 m/s (i.e. within the range of the hydraulic conductivity of the fan deposits mentioned before).

The solute used as a contaminant fluid was NaCl solution with a solute concentration of 0.18 M (10.5 g/l) accompanied with Eosin Y Disodium dye for more accurate visual observation of the contaminant plume evolution. NaCl solution was selected as the contaminant fluid because it gives resistivity values similar to leachate solution.

Experimental procedure

The sand was pluviated into the centrifuge strongbox over the already placed basal resistivity array and tensiometers. The sand was wetted from the bottom of the liner until uniform saturation was achieved. When the model saturation stabilised, the excess water was released and partial saturation achieved within the sand model. After that the valve controlling flow of contaminant was opened. The dyed NaCl solution flowed from the contaminant reservoir to the distributor placed between surface electrodes 13 and 14 allowing infiltration from a 20-mm wide line source, across the model (Fig. 6). During the experiment, 3.8 l of dyed NaCl solution were released into the sand model over a period of 2 h.

The contaminant infiltration experiment took place over 3 days, and the changes in resistivity associated with contaminant plume evolution were imaged in 2-D perpendicular to the line source as a function of time. Resistivity values similar to those derived from the discussed landfill site were observed. Values below 5 Ωm characterise the presence of main contaminant pockets. Values between 5 Ωm and 20 Ωm characterise contaminant dispersion. Values above 20 Ωm characterise the partially saturated sand model.

During the experiment, capillary pressure changes were closely monitored with the situated tensiometers. As a result, the saturation degree of the sand model could be assessed during and after the test.

Discussion of results

The measured resistivity is mostly a function of porosity, of saturation degree (moisture content) and conductivity of the pore fluid (concentration of contaminant). Thus, the main anomaly shown in black colour in Fig. 7 could be related with changes in all three quantities mentioned before. However, as porosity can be considered as being constant and by knowing that uniform saturation was achieved prior to the contaminant release (Fig. 7a); the black anomaly shown in Fig. 7 may be related only to the contaminant distribution within the sand model.

Fig. 7
figure 7figure 7

Tomographies associated with contaminant infiltration experiment

Full size image

As can be seen in Fig. 7b, c, the contaminant was gradually released within the sand model and this resulted in the downward movement of a contaminant plume. The contaminant plume increased in volume and moved towards the base of the model due to the bulk movement of the contaminant solution. In conjunction with advective flow, hydrodynamic dispersion took place in the vertical and horizontal directions. As can be seen in Fig. 7b, c, dispersion occurred outside the main body of the contaminant represented by the 5–20 Ωm resistivity contours.

The water retained initially in the soil in the form of pore and capillary water was displaced gradually by the denser contaminant during and after the release of the total contaminant volume (Fig. 7d–f).

At the end of the experiment, traces of contaminant were retained in the smaller pores of the upper layers of the model by strong capillary forces. However, the main body of the contaminant plume had already reached the base of the model as a result of density-driven migration (Fig. 7f).

Figure 8 shows the plume at the end of the experiment, visible in two longitudinal sections excavated through the model with the observed plume geometry being similar to the final tomography model shown in Fig. 7f.

Fig. 8
figure 8

Geometry of the contaminant plume

Full size image

During the 3-day experiment, the degree of saturation within the sand model, estimated from the tensiometer readings, remained high. In Fig. 7d, 23 h and 33 min after the contaminant release, the degree of saturation was 95–97%. In Fig. 7f, 65 h and 58 min after the contaminant release, the degree of saturation was 85–87%. These values represent the presence of a capillary zone with a continuous liquid phase, so that suction gradients were minimal. This supports the hypothesis that during this test the contaminant evolution was mainly gravity-driven; hence, the discussed model and the apparatus developed may be used for similar bench and centrifuge tests.

Conclusions

Electrical imaging provided details of the internal structure of the discussed waste tip, and confirmed the presence of leachate pockets and plumes. The low resistivity anomalies that extend below the base of the waste tip represent plumes of contaminated fluid, which could migrate towards the underlying water table by the processes of advection flow and hydrodynamic dispersion.

The developed miniaturised electrical imaging equipment and the generic model designed for contaminant infiltration tests were developed on the basis of the field components explored during the site investigation of the waste tip. Comparison of the resulting 2-D tomography images with the observed plume geometry at the end of a 3-day laboratory test showed the electrical imaging equipment to be highly effective even under model conditions. In addition, the developed equipment found to be non-invasive and, therefore, neither disturbed the soil fabric nor compromised the experimental results.

All results suggest that electrical imaging is applicable under both in-situ and model conditions similar to those described in the present paper. Therefore, this method could be applied for similar environmental studies associated with contaminant infiltration. These studies can be investigated in a laboratory environment either in the bench or in the geotechnical centrifuge, provided that contaminant evolution remains gravity-driven.