Model development Directory UMM :Data Elmu:jurnal:J-a:Journal Of Applied Geophysics:Vol44.Issue2-3.2000:

are appreciable density contrasts between the fill and the surrounding geological materials Ž . e.g. Whiteley, 1983; Roberts et al., 1990 , gravimetric techniques could be used for defin- ing the geometry and depth extent of the land- Ž fill. Heterogeneities within the fill e.g. metal . drums may be mapped using magnetometric Ž . Ž and ground probing radar GPR methods e.g. . Hinze et al., 1990 . The shallow seismic reflec- tion and GPR methods are undoubtedly the best Ž geological mapping tools in ideal situations e.g. Hill and Ali, 1988; Ali and Hill, 1991; Beres . and Haeni, 1991 but old landfill sites constitute an entirely different target in which the loose fill is typically associated with poor energy propagation and the clayey horizons limit the usefulness of radar signals. In comparison with the traditional field tech- niques used in natural resources exploration, geoelectrical investigation of landfill sites is on a different scale and requires special attention to details. To start with, the area under investiga- tion is often geometrically constrained and be- deviled by cultural noise, yet the sampling rate and data quality requirements for target defini- tion are more stringent. The consequences of inadequate or unsuccessful site investigation may be more serious in pollution-related stud- ies, with insurance or legal connotations. Thus, in the present economic and legal climate, ade- quate desk-study and model development are essential ingredients for a successful cost-effec- tive geoelectrical investigation of old landfill sites. The main thrust of this paper is to develop a consistent exploration model for anthropogenic deposits in oldrabandoned sites where the nec- essary record of operations is no longer avail- able or never existed, as in the case of unautho- rised dumping grounds. The model will draw from current concepts in geotechnics and con- taminant geochemistry, and stress the complex geometry of landfill sites, the heterogeneous material compositions, and the attendant com- plex biogeomorphic processes in landfill envi- ronments. The analogy between refuse decom- position processes and weathering of geological materials will be explored, and used to guide the development of a conceptual resistivity-vs.- depth model for old covered landfill sites. The versatility of this particular model will be tested using depth soundings from various geographic regions. In line with the exciting geochemical observation that some chemical parameters of Ž landfill leachates vary consistently with age e.g. . Farquhar, 1989; DoE, 1996 , we investigate the relationships between diagnostic leachate chem- Ž ical parameters electrical conductivity, total Ž . . dissolved solids TDS and chloride content and attempt to propose a plausible practical scheme for fill-age and hydrochemical predic- tions using surface geoelectrical measurements. Some illustrative field examples will be pre- sented in support of the proposed predictive scheme.

2. Model development

2.1. Characteristics of landfill sites and anthro- pogenic deposits Landfill sites are geotechnically classified as inert sites, urban and low-strength industrial Ž sites, or high strength industrial or . hazardousrco-disposal sites, depending on the type of fill content. Although there will be no single model that can adequately characterise all types of landfills, we will adopt a generalised approach in this paper. To understand how the attendant processes in landfill environments can influence our geoelectrical measurements, it is instructive to examine the consistent features of models derived from geological, geotechnical, biological and biogeochemical observations on Ž . landfills plus other contaminated lands and rock weathering that can be adopted as the basic building blocks for any geoelectrical model for Ž landfill sites. Three main features complex ge- ometry of landfill sites, heterogeneous material composition, and complex biogeomorphic pro- . cesses in harsh environmental conditions are recognised and adapted into a simplified geo- electric model in this section. 2.2. Geometry of landfill sites Landfill sites come in various shapes, sizes and depths, and as previously mentioned, they may be located in purpose-built facilities, dis- used softrockrhardrock quarries, opencast coal mines or other convenient holes in the ground. They may be situated above, below or astride the regional water table. In some oldraban- doned waste disposal sites, a lining of relatively impermeable material may be present, or the refuse may be in direct contact with granular or Ž crystalline geological materials dubbed the ge- . omatrix in geotechnical parlance . In landfill capping, a soil cover layer is required when returning the site to agricultural or ammenity use and steep sided cover systems are often Ž . incorporated in the cover design see Fig. 1 to Ž maximize the landfill capacity Hall and . Gilchrist, 1995 . The cover system could be multi-layered in sophisticated waste storage fa- cilities or single-layered in some old uncon- trolled landfills. The side slopes of cover sys- tems commonly vary from about 38 up to a maximum of 408. In the UK, for example, the Ž . Fig. 1. Common cover types at old landfill sites. a Ž . Steeply sloping side, and b domed cap. side slopes of the lining system are often re- quired to be in excess of 208 and ‘‘doming’’ of Ž . Ž capping systems Fig. 1 is common Hall and . Gilchrist, 1995 . There may be basal floor slopes to promote leachate drainage to sumps but it can be expected that many pre-regulation landfill sites may have inadequate basal containment and leachate collection systems; there will also be cases where the landfill bottom is neither graded nor lined. Landfills generally range in thickness from about 3 to 20 m but deeper sites Ž . ca. 30 m are known to exist. 2.3. Nature and characteristics of anthro- pogenic deposits Landfill deposits are characterised by com- plex material composition, non-uniform com- paction within each layer, non-uniform decom- position process, non-uniform settlement and Ž . varying pore fluid composition Fang, 1995a . The deposits may be intermixtures of domestic and industrial wastes, soils and exhumed geo- Ž . logical materials see Table 1 but are geotech- nically grouped into four classes: inert wastes, Ž . urban sanitary or compostable wastes, low Ž . strength industrial non-hazardous wastes, and Ž . high strength industrial hazardous wastes. The composition of urban waste will vary from com- munity to community, from country to country, Ž . and from season to season Fang, 1995a as partly illustrated in Table 1. The wastes in old landfill sites may not be as well compacted as in modern regulated landfill practice and will thus have substantial internal permeability. They will, in general, consist of Ž degradable and non-degradable materials food and garden wastes, ashes, paper, textiles, plas- tics, metals, building waste, mill tailings, or- . ganic liquids etc but it is their chemical compo- Ž . sition Table 2 that is important when assessing their potential for groundwater pollution. For example, the chemical composition of the un- processed waste from Zagreb in Croatia listed in Table 1 is 75.8 mineral content and 24.2 Ž . organic matter Kovacic et al., 1995 , and its Table 1 Ž . Ž . Ž . Composition percentage by weight of typical municipal solid wastes. Data taken from Sowers 1968 ; Yamamura 1983 ; Ž . Ž . Ž . Kovacic and Mayer 1995 ; Sanchez-Alciturri et al. 1995a and Wasti 1995 . The superscripts refer to original classifications at data sources Types New York City, Osaka, Zagreb, Meruelo, Ankara, Beijing, USA Japan Croatia Spain Turkey China Food and garden r 19.3 17.7 23.3 52.0 50.8 45 organic wastes Paper, paper board 58.8 37.1 20.5 21.9 8.9 5 Metal 7.6 5.5 2.4 3.2 1.2 1 Glass 8.6 12.3 7.3 4.1 1.4 1 Wood 2.5 2.5 1.1 2.4 – Textiles, rugs 0.8 4.0 3.8 1.4 b Rubber, leather, 0.8 0.3 2.6 1 b Ž . bones , wood Plastic 0.8 15.2 10.8 8.3 2.1 1 Earth, ash, cinder 1.4 32.4 Construction rubble 28.2 1.8 Others 6.7 46 influence on groundwater will be dominated by the amount of heavy metals and trace elements or the organic acid derivable from the solid waste. Many potentially hazardous materials find wide application in industry. For example, poly- Ž . chlorinated biphenyls PCBs have wide indus- trial usage due to their high dielectric constant, fire resistance and thermal stability; they come under various tradenames — for example ‘‘askarel oil’’ found in some old transformers. Some of the common industrial wastes are listed in Table 2. Many of these wastes ultimately find their resting places in landfill sites. Methylene Ž . chloride, trichloroethylene TCE , toluene and m-xylene are among the hazardous organic compounds most commonly found in landfill sites. Since landfills are a complex mixture of an- thropogenic deposits, their physical properties would show a wide range of variation. For example, the density of urban refuse varies from 275 kgrm 3 to 6400 kgrm 3 depending on the amount of metal and construction debris it con- Ž tains e.g. Knight, 1990; Sharma et al., 1990; . Fang, 1995a; Sanchez-Alciturri et al., 1995b but in general, old landfills normally have low Ž density and seismic velocity Whiteley and Jew- . ell, 1992 . The published resistivity of solid waste and contaminated substrate range from Ž 1.5 to ca. 20 V m e.g. Knight et al., 1978; Laine et al., 1982; Everett et al., 1984; Carpen- . ter et al., 1991 with the associated leachate Ž being highly conductive Whiteley and Jewell, . 1992 . The seismic P-wave velocity of urban refuse may range from 180 to over 700 mrs Že.g. Knight et al., 1978; Calkin, 1989; Sharma . et al., 1990 depending on the degree of com- paction and state of saturation with the cover material having higher velocities than the under- Ž lying fill e.g. Calkin obtained a cover velocity of 320 mrs and refuse velocity of 180 mrs . over a municipal landfill near Chicago . Fortu- nately, these physical properties are often markedly different from those of the host geo- logical materials so that geophysical methods can be used for determining the volumetric distribution of fill. 2.4. Leachate formation and dispersal Degradable refuse will decompose with time into organic andror inorganic soils plus other byproducts depending on its chemical composi- Table 2 Ž . Some industrial wastes commonly buried in landfill sites adapted from Fang, 1995a with additions Sources Type of wastes Approximate composition Food and food products Additives Trimmings Organics and acids Grain mills Residues Meatrfish Sludges Paper and paper products Sludges, pulping Sulphates, organics, sludges soaps, mercaptans Soaps, detergents Sludges Surfactants, polyphosphates, aluminium–copper–oxides Textile products: silk,cotton, Sludges Acids, alkalis, metallic wool, synthetics salts, solvents Leather products Sludges Chrome salts, oils, dyes Wood and wood products Sludges, residuesr Solvents, preservatives mill tailings Paints, varnishes Sludges Metallic salts, toxic liquids Energy and petroleum Sludges, residues, Hydrocarbons, acids, coal, nuclear and cindered coking, metallic salts, petroleum refining fly ash radioactive materials Metals, fabricatedr Sludges, slag, Sulphur, chlorides, scrap metals slime, tailings phenols, PCBs, oils, scrap heap grease, chrome, alkalis, acids, metallic salts Miningrmineral processing Sludges, mine Acids, cynanide, tailings metallic salts Chemicals, fertilizer Sludges Sulphuric acids, organo-phosphates, copper sulphate, mercury arsenates tion. Plausible biogeomorphic processes are summarised here and draws from observations and concepts in other geoscientific disciplines Že.g. Schoell, 1980; Levinson, 1980; Perec, 1981; Robinson et al., 1982; Crawford and Smith, 1984; Jones, 1985; Bennet and Siegel, 1987; Palacky, 1987; Farquhar, 1989; Robinson, 1989; Yong et al., 1992; Bell and Jermy, 1995; . Fang, 1995a,b . 2.5. Mechanical decomposition Urban waste deposit undergoes an initial short-term process of mechanical alteration due to loading and its bulk density and other physi- cal properties may change in response to this Ž . preliminary settlement process Fang, 1995a . Ž Other transient mechanical loads such as snow, . rainwater and surcharge loads will also con- tribute to the settlement process. Owing to the heterogeneous nature of the usually organic-rich urban waste, the distribution of settlement will Ž . be non-uniform Fang, 1995a and often leads to severe fracturing of the top seal of the landfill cover. The top seal is then highly vulnerable to erosion and infiltration of rainwater and snow- melts. 2.6. Physico-chemical and microbial weather- ing The decomposition of landfilled wastes by Ž long-term physico-chemical, chemical notably hydrolysis, hydration, carbonation, oxidation and . Ž solution and biological degradation mostly mi- . crobial processes cause the dissolution or dete- rioration of landfill materials, gas generation and production of leachate. Landfill sites pro- vide ideal environments for bacterial colonies to Ž grow since nutrients availability, and substrate composition and temperature requirements are . met and most bacteria flourish in the aerobic Ž condition above the groundwater table e.g. . Fang, 1995b . Initially, the microbial degrada- tion of landfill materials occurs under aerobic conditions. As the oxygen becomes depleted by the microbial activity, anaerobic conditions rapidly set in and the biodegradation of organic materials becomes anaerobic. Methane gas is generated bacterially from the abundant organic materials under the prevailing anaerobic condi- tions. Although it is now widely required in land- Ž filling that refuse be dynamically compacted to . minimise volume and then covered by a thin Ž layer of soil or clayey seal to minimize wind . dispersion at the end of each operative day, rainwater ingress commonly takes place in wet seasons during landfilling and for capped sites, through cracks in the cover and along contact points between fill and host material. Infiltrating rainwater, groundwater, or other liquids dis- posed of within the wastes will dissolve some soluble mineral constituents of the landfill once the absorbent or field capacity of the fill is exceeded and free drainage of water can occur. This leaching process may remove the common mineral elements such as calcium, magnesium, potasium, nitrogen and phosporous or remove Ž . the bonding materials e.g. clay resulting in changes in matrix cement or the ion concentra- tion within the landfill-porewater system and consequently causing significant physical prop- erty changes. The ion exchange reaction, speeded up by bacterial activity, also causes changes to the structure and composition of the Ž landfill–porewater system. For example, hy- drogen sulphide may be converted to sulphuric acid, methane gas may become converted to carbon dioxide and water by microbial metabolism while the carbon dioxide by-product . may combine with water to form carbonic acid . Ž . The resulting liquid termed leachate is rich in fungi, bacteria, inorganic salts and organic mat- ter; but the compositional trend may be water and dissolved inorganic salts, water and dis- solved organic wastes and organic fluids, or simply organic acids depending on the availabil- ity of solvent and solute types in the leached mass. In terms of availability of solute types, the bottom ash residues from urban solid waste incinerators contain considerable amounts of Ž leachable heavy metals lead, zinc, cadmium, . Ž copper, and chromium and salts Bahout et al., . 1995 , and this will be reflected in the composi- tion of leachate derived from such deposits. If pyrite is available in wastes containing mine tailings, it may be oxidised to sulphuric acid while the decomposition of vegetation may pro- duce organic acids. PCBs have low solubility in water, and if present in the landfilled wastes, we can expect low concentrations of PCBs in the leachate dissolved phase, but they are more likely to occur in solid organic matter or in the oil fraction. Since higher volumes of water would have passed through poorly compacted waste materi- als in old landfill sites compared to modern compacted landfills, there should be relatively lower concentrations of chemical constituents in leachates derived from old landfills with effi- Ž . cient migration processes Radnoff et al., 1992 . Note that for poorly consolidated or uncom- pacted wastes at shallow burial depths, and with inadequate capping, the absorbent or field ca- pacity may be achieved within only a few years following initial waste emplacement, thus al- lowing early generation of leachate. For well compacted and deeply buried wastes with low permeability capping, the field capacity may not be achieved until after several years, but the waste compression process may facilitate early formation and expulsion of leachate from the Ž lower layers of waste into the substrate Lewin . et al., 1997 . Note also that the decomposition of organic and inorganic solids will be associ- Ž ated with a fill volume change causing shrink- . age or swelling . In general, the pore fluids produced from landfill are mostly acidic, but will vary in com- Ž . position from country to country Table 3 , community to community and with season. Bac- teriogenic methane is formed by fermentation of organic material in conditions of depleted oxy- gen supply, and for a given leachate, a number of chemical changes occur as it evolves from Ž . the aerobic acetogenic stage to the anaerobic Ž . Ž . methanogenic stage Robinson, 1989 ; the to- Ž . tal organic content TOC , total free fatty acids content or acetone content and TDS are high during acetogenesis and low during methano- genesis. That is, the composition of the leachate will change as the refuse in the landfill ages Ž . Farquhar, 1989 . In general, the leachate from a young landfill may be characterised by high levels of organic acids, ammonia and TDS, but as much of the biodegradable mass is broken down with time, the concentrations of these parameters will decrease in the leachate pro- Ž . duced from the ageing landfill see Table 4 . The leachate may contain toxic or hazardous substances in solid or gaseous forms and might show up as high concentrations of chloride, iron and zinc ions. Those elements with high ionic mobility generally have the highest concentra- tion whilst those having low mobility usually have the lowest concentration in leachates Ž . Bagchi, 1987 . The pH tends to increase with Ž time i.e., from an initial acidic state to a neutral . state while the biological and chemical oxygen Ž . Ž demands BODrCOD decrease with age see . Table 4 . The concentration of organic carbon Table 3 Ž . Ž . Regional variations in leachate composition. Data for columns 2–6 are from Robinson et al. 1982 , Ehrig 1983 , Fang Ž . Ž . Ž . Ž 1995a , Niininen et al. 1995 and Vendrame and Pinho 1997 , respectively. All quantities are in mgrl except pH in . standard units . N s nitrogen. P s Phosphorus. TDS s total dissolved solids. BOD s biological oxygen demand. COD s chemical oxygen demand. TOC s total organic content Parameter England Germany USA Finland Brazil TDS 1–520 402–6794 pH 6.2–7.4 6.1–8.0 3.7–8.5 3–8 6.05–7.51 BOD - 2–8000 180–13 000 22 000–30 000 1–3900 COD 66–11 600 3000–22 000 800–50 000 52–5200 90–2000 TOC 21–4400 17–1900 Ammoniacal N 5–730 741 0.3–480 1–480 14–1080 Total P or phosphate - 0.02–3.4 5.7 0.5–130 - 0.02–3.9 0.057–2.312 Chloride 70–2777 2119 50–2400 1–600 275–1949 Sulphate 20–750 Calcium 165–1150 80–1300 240–2400 65–69 Sodium 85–3800 Potassium 28–1700 Magnesium 12–480 250–600 64–410 23.8–59.8 Iron 0.1–380 15–925 0.15–1640 - 0.1–692 4.504–9.9 Manganese 0.3–26.5 0.7–24 0.1–7.4 Zinc - 0.1–1.0 0.6–5.6 0.02–130 - 0.1–1.4 Cadmium - 0.005–0.01 0.0052 - 0.0001–0.0047 0–0.033 Chromium - 0.05–0.16 0.275 - 0.001–0.134 0.005–0.056 Nickel 0.05–0.16 0.166 0.15–0.9 - 0.003–0.394 0–0.333 Lead 0.05–0.22 0.087 - 0.001–0.042 0.021–0.7 Copper 0.01–0.15 0.065 - 0.001–0.056 0.008–0.19 Table 4 Ž . Typical changes in leachate concentrations with age of refuse after Farquhar, 1989; Birks and Eyles, 1997 . All quantities Ž . are in mgrl except pH in standard units Parameter Age of refuse 0–5 years 5–10 years 10–20 years 20 years TDS 10 000–25 000 5000–10 000 2000–5000 - 1000 pH 5–6 6–7 7–7.5 7.5 BOD 10 000–25 000 1000–4000 50–100 - 50 COD 15 000–40 000 10 000–20 000 1000–5000 - 1000 Ammoniacal N 500–1500 300–500 50–200 - 30 Total P 100–300 10–100 - 10 Chloride 1000–3000 500–2000 100–500 - 100 Sulphate 500–2000 200–1000 50–200 - 50 Calcium 2000–4000 500–2000 300–500 - 500 Sodium q potassium 2000–4000 500–1500 100–500 - 100 Magnesium q iron 500–1500 500–1000 100–500 - 100 Zinc q aluminium 100–200 50–100 10–50 - 10 Alkalinity 10 000–15 000 1000–6000 500–2000 - 500 often exceeds 8000 mgrl in the leachates from young landfills and 465 mgrl in leachates from Ž . old landfills Dearlove, 1995 . The composition of the leachate will depend on the type and age of fill, water infiltration rate Ž . and pH Farquhar, 1989 but the rate and quan- tity of leachate and landfill gas production will be affected by the depth of burial of fill, re- gional climatic conditions, variations in water table, the landfill capping practice and fluid inflow and outflow controls at the site. It can be Ž expected that surface layers of refuse i.e., shal- . low burial depth may experience rapid aerobic decomposition whilst the bulk of the waste at depth may have only been partially decomposed under anaerobic conditions thus leading to dif- ferent physical properties. Also, landfill degra- dation will be quicker in humid tropical regions than in cold regions and for identical fill com- position and water influx–efflux conditions, the leachate from the warmer climate will have a higher concentration of dissolved materials rela- tive to the background groundwater composition Žin analogy to variation in saprolite composition Ž .. in chemically weathered rocks Palacky, 1987 as also suggested by the data presented in Table 5. If precipitation is the main source of water for leachate generation in the fill, then it can be expected that storms will have a role in leachate discharge. The transport of leachate through the landfill is slow, unsteady, non-uniform and sometimes Ž . discontinuous Fang, 1995a depending on the degree of compaction of the fill and seasonal changes in water supply to the system. Within the landfill, this liquid may collect in various Ž . areas e.g. perched saturated zones or mound at the bottom of the landfill. This leachate starts seeping as soon as enough hydrostatic head is developed. Biochemically controlled exothermic reactions are known to cause higher groundwa- Ž ter temperatures in leachate MacFarlane et al., . 1983 and because of the ingress of leachate from the upper leached zones, the temperature in the lower portion of the landfill is often significantly higher than elsewhere in the Ž . leached section Fang, 1995a . Consequently, there are higher bacterial activities and higher ion exchange reactions in the lower parts of the landfill as time progresses. These microbial– chemical decomposition reactions may cause significant changes to the existing pore fluids Žand to the substrate if the fill is in direct . contact with geological materials . M.A. Meju r Journal of Applied Geophysics 44 2000 115 – 150 124 Table 5 Ž . Comparison of some leachate components at four sites in different geographic regions. Data for the Durban sites IP10-SS2 South Africa are from Bell and Jermy Ž . Ž . Ž . Ž . 1995 ; SS2 is a surface stream sample and the rest are from inspection pits IP . Data for site NE northern England are from Kalteziotis et al. 1995 . Data for Ž . Ž . Ž . Ž . Ž . Beverly BEV and Upper Ottawa Street UOS landfill sites Canada are from Birks and Eyles 1997 . Electrical conductivity EC is in mSrm and except for Ž . pH, the rest are in mgrl. In the last two columns, the average values of the parameters as well as their respective ranges in parenthesis are given Parameter Sites IP10 IP11 IP17 IP20 IP9 IP4 IP19 IP21 SS2 NE BEV UOS Ž . pH 6.9 7.5 6.9 6.9 7.3 7.45 7.7 7.05 7.35 7.2 7.9 7.6–8.3 Ž . Ž . TDS 11047 9902 557 2166 9726 9003 3722 2807 261 933 40–1764 8181.2 6140–10108 Ž . Ž . EC 1397.4 1530 85.7 300.9 1448.4 1254.6 581.4 418.2 38.8 501 140.1 19–349.9 1123.3 750–1500 Ž . Ž . Chloride 3938 3832 76 504 3782 3620 1033 820 71 510 150 0–750 1344 29–3150 Ž . Ž . COD 656 1920 6240 272 3520 2960 960 448 40 800 97 0–283 1772 96–4706 On passing through the base of the landfill, the metal ions in solution may be removed from the aqueous phase by ion exchange, sorption or Ž precipitation onto the substrate especially if . clayey . However, organic carbon in colloidal form in the leachate often has higher cation exchange capacity than clay and can sorb high Ž concentrations of metal ions from solution as can some inorganic colloids which form under . certain chemical conditions . The metal ions sorbed preferentially onto the surface of col- loidal particles may thus by-pass the natural attenuation processses as the leachate seeps into Ž . the substrate Dearlove, 1995 . Within the sub- strate, it mixes with groundwater forming a leachate plume. Initially, on entering the anaerobic groundwa- ter system, the organic material in the leachate is slowly biodegraded forming more acids which Ž may react with aquifer materials cf. Bennett . and Siegel, 1987 with attendant changes in the fluid chemistry near the water table. In this deoxygenated environment, inorganic materials Ž . in the leachate e.g. iron, manganese may be dissolved in the groundwater. The dispersing leachate extends laterally and vertically as it sinks towards the bottom of the substrate form- ing a 3-D contaminant plume that may be steeply Ž . dipping see Fig. 2 . The amount of groundwa- ter contamination resulting from this invasion will depend on the hydrogeology of the area and the attenuation capacity of the substrate. It may be effectively diluted and dispersed by ground- water in highly permeable geological formations with high flow rates. Given enough time in less permeable formations, or with slowly moving Ž groundwater, the plume laden with inorganic . salts may enhance mineralisation of groundwa- ter. Since it is a moving and continuously evolv- ing 3-D feature, it will in time be dispersed over a sizeable area, possibly with distinct composi- tional zonations. 2.7. Conceptual resistiÕity model A conceptual resistivity model can be devel- oped in line with the above biogeomorphic and hydrochemical considerations, since the observ- able geoelectrical response of landfill sites and environs will vary in relation to significant changes in the chemistry of subsurface pore fluids. It can be expected that surface layers of refuse may experience rapid aerobic decomposi- tion whilst the bulk of the waste at depth may have only been partially decomposed under anaerobic conditions, thus leading to different physical properties. For a leachate-generating landfill in contact with granular substrate, the interactions between the invading leachate and substrate material may cause geochemical alter- ations of substrate depending on its buffering and cation exchange capacities. Fig. 2. Geometry of typical 3-D contaminant plume migrating downgradient from a landfill site. Accordingly, several zones are recognised in the generalised conceptual model shown in Fig. Ž . 3. The top soil and clay cap i.e., zones 1 and 2 form the uppermost confining layers. The resis- tivity of the top soil will vary from region to region and with season but the usually 0.3 to 1 m thick clay cap will in general be relatively conductive. Landfill sites are notorious for high Ž levels of soil gas principally methane and car- . bon dioxide . Fractured clay caps would also Fig. 3. A conceptual resistivity model for old landfill sites with leachate generation and migration into groundwater system Ž . in granular substrate and environs. Part of the sub-water table hydrogeological processes were adapted from DoE 1996 . allow gas migration. The geoelectrical signa- tures of zones 1 and 2 will be affected by seasonal changes. In wet seasons, infiltrating water may drive out gas thus lowering its bulk resistivity. Conversely, rising vapours in the dry periods will drive out soil moisture with atten- dant increase in bulk resistivity. The degree to which these changes are detectable will depend on the nature of the materials in zones 1 and 2. For instance, the influx of gas into pore spaces in some clays may cause appreciable resistivity Ž . increases if water or leachate is displaced but not so for a highly resistive dry sandy topmost layer. If these cover materials are iron-rich, it is conceivable that the rising methane and sulphur dioxide in landfills may lead to the formation of a pyritic geochemical alteration halo akin to the so-called ‘sulphide chimney’ over fractured hy- drocarbon accumulations overlain by iron-rich Ž sediments or redbeds e.g. Oehler and Stern- . berg, 1982,1984; Ostrander et al., 1983 which might be detectable in 3-D or time-lapse geo- electrical surveys. The clay cap is underlain by a zone of perva- sive leaching of refuse and residual products Ž . i.e., zone 3 in Fig. 3 . This is the top part of the landfill waste where oxygen and bacterial sup- Ž ply is abundant and will thus decompose . quicker than the deeper parts . The bulk resistiv- ity of this oxidized zone will show a relative Ž increase with time as the organics for microbial degradation become depleted in supply, oxida- tion of inorganics tends towards completion and much of the soluble elements have been re- . moved . The leaching of clay minerals from the original landfill-soil mixture would leave behind the non-degradable fill material and siliceous geomaterial with average soilrrock grains of Ž . larger size fraction cf. Witmer et al., 1984 and of relatively higher resistivity than the parent landfill material. It may thus be relatively resis- tive in comparison with the clay cap and the underlying zone of incomplete refuse decompo- Ž . sition i.e., zone 4 . However, the rate of water flux through zone 3 will affect its eventual resistivity characteristics — it will be more resistive for fast flows in comparison to slow fluxes that will cause lower resistivities. Zone 4 occurs in the lower part of the waste deposit and is dominated by relatively immature leaching and will exhibit relatively low bulk resistivities. For landfill resting on highly im- permeable substrate, leachate mounding may occur in the basal part of the waste deposits rendering it the most electrically conductive part of the entire landfilled section. However, if the landfill rests directly on a relatively permeable substrate saturated with groundwater, fluctua- tions in the water table may occur within a section encompassing the basal part of the land- Ž . fill zone 5 and the uppermost part of the Ž . substrate zone 6 as schematised in Fig. 3. The mineral salts and organic material leached from the fill materials will be deposited near the water table depending on the local hydrogeolog- Ž ical and Eh–pH conditions in analogy with the well-known process of supergene enrichment of Ž . metallic sulphide minerals Levinson, 1980 and . chemical weathering of soilsrrocks causing an Ž increase in TDS and therefore electrical con- . ductivity and other chemical parameters in the Ž . porewater cf. Knight et al., 1978, Fig. 8 . Due to water table fluctuations, the zone of deposi- tion or mineral enrichment may extend from the basal part of the landfill into the upper part of the subjacent geological formation or may lie well beneath the base of the landfill depending on the permeability, fluid saturation, groundwa- ter flow and dispersion characteristics of the substrate and the mobilities of the ions in solu- tion. Below the water table, the leachate mixes with groundwater and reacts with substrate ma- Ž . terial cf. Bennet and Seigel, 1987 forming a Ž . relatively conductive plume zone 7 . Beyond Ž . the plume, the uninvaded substrate zone 8 will have TDS and conductivity parameter values Ž intrinsic to the natural medium i.e., background . concentrations . The main tenet of this hypothesis is that the TDS and conductivity profiles will peak near the water table and tail-off in either direction outside the zone of mineral enrichment. It can be expected that this zone will have a dis- cernible conductivity signature on geoelectrical soundings in leachate-bearing landfill environ- ments. For unsaturated substrate in contact with leachate-generating landfill deposits, the leachate will advance with the infiltrating water and displace the air or residual connate water in the pores spaces of the substrate. This displace- ment will cause a decrease in the resistivity of sandy or carbonate formations but the effects may not be geoelectrically appreciable in some clayey substrate. If the unsaturated section were thick enough, the rate of advance of the leachate Ž front and hence the depth location of the zone . of maximum leachate concentration will de- pend on the buffering and cation exchange ca- pacities of the substrate. For instance, calcite- rich sands may attenuate the leachate better than Ž . dominantly siliceous sands e.g. DoE, 1996 with consequent differences in the position of occurrence of leachate-related geoelectrical con- ductors in unsaturated substrates. In Fig. 4 are shown the typical landfill sound- ing curves from various geological, climatic and geoenvironmental settings. Note that all the curves are of the minimum or H-type. The most conductive segment of each sounding curve cor- responds to the saturated basal section of the respective landfill andror leachate-invaded sub- strate. Note, however, that an apparently highly resistive terminal segment of a sounding curve Ž at a landfill site as exhibited by the curves . shown in these figures may not always be indicative of a resistive nature for the substrate but rather may be due to 3D distorting influ- ences of the typically complex site geometry Ž . Ž Fig. 4. Typical resistivity sounding curves from landfill sites in different geographical regions. a Australia Knight et al., . Ž . Ž . Ž . Ž . Ž . Ž . 1978 , b USA Carpenter et al., 1991 , c UK Barker, 1990 , and d UK Meju, 1995a . especially when the expanding electrodes oc- cupy positions outside the actual confines of the landfill site. 2.8. Relationship between geoelectrically impor- tant hydrochemical parameters Electrical conductivity is usually taken as a measure of the total dissolved salts in ground- water by hydrogeochemists whereas chloride content is used as a conservative leachate indi- cator parameter in water sample studies since, apart from dilution, it undergoes very little chemical or biological change in the groundwa- Ž . ter system Baedecker and Apgar, 1984 . Some chemical analyses of groundwater contaminated by leachate seeping from a landfill site that Ž closed in 1989 in Durban, South Africa Bell . Ž and Jermy, 1995 are shown in Table 5. The data were presented at the GREEN ’93 confer- ence in June 1993 and were probably collected . in 1992 . Also shown in this table are the chemical parameters for leachate from a 6-year- Ž old landfill in northern England Kalteziotis et . al., 1995 and for the leachates from the Beverly Ž . landfill active from 1965–1980 and Upper Ž . Ottawa Street landfill active from 1950–1980 Ž near Hamilton in Ontario, Canada Birks and . Eyles, 1997 . For the Canadian examples, the Ž . average, minimum and maximum values of the respective parameters for the monitoring period 1980–1994 are presented. The levels of TDS and conductivity for the leachate-impacted groundwater from downgradient measurements in the vicinity of the Durban landfill are higher than those recorded within the younger landfill site in the UK and comparable to those for the Canadian sites. Since even higher values may be obtained for leachate samples from the Dur- Ž ban landfill than those presented here possibly . diluted offsite samples , the data in Table 5 could be interpreted as supporting the earlier contention that apart from age, climatic differ- ences might lead to variations in the amount of soluble material removed from a given landfill material in different geographical locations. The spatial variation of plume properties is of geoexploration interest and will be examined here using the data for Durban landfill. The Durban landfill sampling points were distributed over a grid across the southeastern border of the Ž landfill where seepage was occurring see Bell . and Jermy, 1995, Fig. 3 . Sampling stations IP9, IP4, IP19 and IP21 in Table 5 were located along the axis of the plume at increasing dis- tances from the southeastern border of the Dur- Ž ban landfill the typical inspection pit spacing . was about 35 m . For these stations, note the decrease in TDS and electrical conductivity with increasing distance from the landfill. To illus- trate the relationship between groundwater con- Ž . ductivity and TDS and chloride content in the plume emanating from the Durban landfill, the data from all the sampling stations in Table 5 are shown graphically in Figs. 5 and 6. Note the almost linear relationship between conductivity and TDS in Fig. 5 and between conductivity and chloride content in Fig. 6. Simple straight- line fitting of the Durban data yielded the rela- Fig. 5. Relationship between fluid conductivity and TDS. The round symbols represent the data from Durban landfill Ž . in South Africa Bell and Jermy, 1995 . The data from Beverley and Upper Ottawa Street landfill sites in Hamil- Ž . ton, Canada Birks and Eyles, 1997 are shown by triangu- Ž . lar symbols. The data from NE Brazil Meju et al., 1997 are shown by cross symbols. Fig. 6. Relationship between conductivity and chloride content for two landfill sites. The data from the Durban site are shown by circular symbols while the triangular symbols represent the data from Morley landfill site in Ž . Australia Buselli et al., 1990 . tionships TDS s y54.4 q 7.04s and Cl s w y257.2 q 2.83s , where the electrical conduc- w Ž . tivity of the groundwater s is in mSrm and w the other two parameters are in mgrl. However, it is often preferable to consider the relation- Ž ships between the logarithms of these data cf. . Table 6 . The resulting relationships for the same data sets are log TDS s 0.8 q 1.015 log s w = or TDS s 6.3096 s 1.015 1 Ž . Ž . w log Cl s y0.256 q 1.2 log s w = or Cl s 0.5546 s 1.2 . 2 Ž . Ž . w Some published data for the Beverly and Upper Ottawa Street landfills in Hamilton in Ž . Canada Birks and Eyles, 1997 and Morley Ž . landfill site in Australia Buselli et al., 1990 are also presented in Figs. 5 and 6 for comparison with the Durban data. The fluid conductivity and chloride content data from Morley landfill Ž show the same trend as the Durban data see . Fig. 6 . Although not an ideal data set, the published average, minimum and maximum conductivity and TDS values for the two landfill Ž . sites in Hamilton Birks and Eyles, 1997 are Ž . also in accord with the Durban see Fig. 5 . It is Ž . interesting as shown in Fig. 5 and Table 6 that a similar linear trend exists in data from a non-landfill environment — a hydrogeochemi- cally zoned regional aquifer with high back- ground inorganic salts near the city of Picos in Ž . northeast Brazil Meju et al., 1997 , which could suggest that the leachates from the selected landfill sites derive their conductive nature largely from inorganic salts. It might be expected that if similar relation- ships exist between bulk resistivity and TDS, then it will be possible to predict compositional trends in hydrochemical parameters such as TDS and chloride content from surface geoelectrical Ž . soundings e.g. Buselli et al., 1990 . No such relations have been found as yet for routine geoelectrical use. However, from a dc resistivity study of acid-mine drainage problem, Ebraheem Ž . et al. 1990 obtained the relation, log s s y0.333 q 0.6453 log TDS Ž . b = or s s 0.4645 TDS 0.6453 3 Ž . Ž . b while a re-appraisal, by this author, of the re- sults of a TEM–AMT study of part of a highly saline, relatively homogeneous, sandstone Table 6 Summary statistics for regression analysis of fluid conductivity vs. TDS in some groundwater systems. In columns 2 to 4, the top set of numbers relates to logarithmic fitting while the numbers in brackets are for non-logarithmic fitting. The data Ž . used were from various sources Bell and Jermy, 1995; Birks and Eyles, 1997; Meju et al., 1997 Ž . Regression parameter Durban, S. Africa Picos, NE Brazil Combined set inc. Canada Ž . Ž . Ž . Intercept 0.8 0.05 y54.4 341 1.1 0.17 81.6 60.9 0.8 0.09 7.69 88 Ž . Ž . Ž . Slope 1.015 0.003 7.04 0.21 0.9 0.01 6.68 0.3 1.018 0.009 7.01 0.11 Ž . aquifer Meju et al., 1997 suggested the rela- tion log s s y0.3215 q 0.7093 log TDS Ž . b = or s s 0.477 TDS 0.7093 4 Ž . Ž . b where s is the bulk conductivity of the forma- b tion in mSrm and the TDS is in mgrl. For the sake of completeness, we may tenta- Ž Ž . tively adopt the above relations Eqs. 3 and Ž .. 4 in the present discussions but note that they were determined for relatively homogeneous materials and thus, may not hold in the typical landfill environment where significant quantities of conductive metal and clay may be present. Moreover, even under favourable conditions, Ž . the two constants intercept and slope in the Ž . Ž . above Eqs. 3 and 4 may have to be evaluated for different landfill environments to enable a working relation between the bulk conductivity of the contaminated homogeneous substrate, or saturated basal fill and the fluid conductivity, to be determined in conjuction with variants of the other relations given above. Assuming that Eq. Ž . 4 is applicable to landfill environments, com- Ž . Ž . bining Eqs. 1 and 4 leads to log s s 0.2452 q 0.7196 log s b w = or s s 1.7587 s 0.7196 5 Ž . Ž . b w relating fluid and saturated fill conductivities. An alternative approach may be sought based Ž . on the popular Archie’s 1942 law, s s b ks f m , where f is porosity, m is the cementa- w tion factor and k is a constant. Along this line, Ž . Yaramanci 1994 developed a relation between in situ resistivity and water content in salt rocks; a simplification of the formula yields the work- ing relation s s s W m 6 Ž . b w with W m as the water content and m 1.6–1.9. Ž This would suggest that under ideal i.e., clay- . free conditions, we may directly estimate the proportion of contaminated water in an invaded homogeneous geomatrix of known bulk and Ž . fluid conductivities. It is stressed that Eq. 6 is adopted here only for illustration purposes-the Ž reader is referred to Worthington 1993 and . references therein for sophisticated adaptations of Archie’s law. 2.9. Conceptual prediction of hydrochemical parameters and age of saturated fill The chemical parameters and age of fill are important components of any genetic recon- struction scheme. Variations in fill compositions of different ages have been noted by several Ž . workers see, e.g. Knight et al., 1978 . How- ever, from a geoelectrical viewpoint, an exciting development in geochemical characterisation of landfill leachates is the observation that these Ž variations are consistent with age e.g. Far- . quhar, 1989; DoE, 1996 as can be gleaned from Table 4. If the approximate concentrations of the relevant hydrochemical parameters can be predicted by virtue of the emprical relationships developed in the previous section, it follows that one can, at least at a conceptual level, predict the age range of a given saturated fill using the information furnished by surface andror borehole geoelectrical measurements. The main problem that will bedevil such an approach is the fact that there are three main sources that contribute to the observed leachate composition — infiltrating groundwater or rain, waste deposits and ambient geological materi- als. Section 3 will focus on the effective integra- tion of the above concepts and models in rou- tine geoelectrical investigations.

3. Development of a consistent investigative geoelectrical approach