Biogeophysics for Optimized Characterization of Petroleum-Contaminated Sites

Abstract

Oil spills are common occurrences on land and in coastal environments. To remediate oil spills, the contaminated volume has to be defined, appropriate remedial measures should be undertaken, and evidence must be provided for the successful remediation. Geophysical techniques can aid site investigation and remediation efforts. The insulating properties of hydrocarbons make them ideal targets for employing a variety of geophysical techniques for their characterization. Nonetheless, the geophysical response of hydrocarbon-contaminated sites is non-unique and depends on factors such as: (1) the release history, (2) hydrocarbon distribution and partitioning into different phases (vapor, free, dissolved, entrapped, and residual phases) in the unsaturated and saturated zones, (3) seasonal hydrologic processes, (4) extent of biodegradation and (5) aquifer salinity and host lithology. Where the contaminants have been biodegraded, the geophysical response depends on the by-products of different terminal electron acceptor processes (TEAPs). In this chapter, we review the different pathways by which TEAPs mediate geophysical property changes. We provide select field case studies from hydrocarbon-contaminated sites across the globe, including different climatic regimes and water salinity conditions. We show that the geophysical response can be transient, hence, data interpretation should be guided by an understanding of the hydrobiogeochemical processes at each site.

9.1 Introduction

Contamination of soil and sediments in the unsaturated and saturated zones by accidental oil spills (e.g., from pipeline breakage, filling stations, refineries, tankers, well blowouts, etc.) is a common occurrence. Since 2010, close to 100 major spills have been reported with an estimated minimum of 5.5 million barrels released into the environment (https://en.wikipedia.org/wiki/List_of_oil_spills#Complete, access 1/25/2022), making hydrocarbons, especially light non-aqueous phase liquids (LNAPL), a major source of groundwater contamination. Both natural attenuation and active remediation measures are used to mitigate and clean up hydrocarbon-contaminated environments (Brown and Ulrich 2014). Geophysical techniques play a crucial role in mitigation and cleanup efforts. Geophysical techniques are used to aid in contaminant delineation, estimate the impacted volume of the subsurface, characterize the site geology/lithology, delineate preferential contaminant flow pathways, as well as monitor the efficacy of the remediation process (Atekwana and Atekwana 2010). Geophysical techniques have the added advantage of being relatively inexpensive, minimally invasive to non-invasive, and provide continuous data coverage over spatial and temporal scales not captured by other more invasive techniques commonly applied at LNAPL-contaminated sites.

Geophysical techniques measure the contrast in the physical properties between a target of interest and the background. LNAPLs and other petroleum products are resistive, and hence a fresh oil spill should result in a resistive response as the LNAPLs replace more conductive pore water. Indeed, fresh spill experiments document the resistive signatures of LNAPLs (DeRyck et al. 1993). The resistive signature was the model used for investigating hydrocarbon-contaminated sites until the seminal work of Sauck et al. (1998), where the authors documented conductive responses and provided strong evidence linking geophysical signature changes to microbial processes at a jet fuel spill site. Today, a variety of geophysical techniques such as electrical resistivity imaging (ERI), electromagnetic (EM) low frequency methods, ground penetrating radar (GPR), induced polarization (IP), and self-potential (SP) have been extensively used to investigate hydrocarbon-contaminated field sites (Atekwana and Atekwana 2010; Abbas et al. 2017, 2018 and references therein). As in other bioreactor systems, the geophysical techniques measure the changes in physical properties resulting from the intermediate and end products of different microbial-mediated processes occurring during the breakdown of hydrocarbons. Although fresh spills occur regularly, in this chapter, we focus on the use of geophysical characterization of aged (several years to decades old) spills in the subsurface, as thousands of such sites exist globally, and need to be remediated. The choice of geophysical methodologies to be used at field sites requires an understanding of the geophysical properties measured by each methodology.

In this chapter, we provide an overview of the terminal electron acceptor processes (TEAPs) at hydrocarbon-contaminated sites. Then, we discuss pathways by which these TEAPs and their by-products mediate changes in physical properties of sedimentary systems. Finally, we provide select case studies of how the by-products of the TEAPs are expressed in geophysical signatures, and thus their utility in investigating LNAPL impacted sites in different environments.

9.1.1 Terminal Electron Acceptor Processes at LNAPL Impacted Sites

When bacteria degrade highly reduced petroleum hydrocarbons, they must contend with a large number of electrons. Most bacteria contain proteins and other compounds in their cell membranes that facilitate the transport of electrons, constituting an electron transport chain (Madigan et al. 2018). An electron transport chain allows bacteria to acquire energy, because as electrons are transported through the chain, protons are exported, and a membrane potential is created. The membrane potential can be converted into mechanical and chemical energy.

The greatest amount of energy is derived if oxygen is used as a terminal electron acceptor in aerobic respiration. Because of their much longer evolutionary history compared to animals and plants, many bacteria are endowed with proteins that can relay the electrons not only onto oxygen, but also onto other dissolved and solid terminal electron acceptors. These other electron acceptors include soluble nitrate and sulfate ions, as well as minerals containing oxidized iron or manganese. Based on thermodynamic considerations, the use of electron acceptors other than oxygen results in less energy gain for the bacteria. If there are no external electron acceptors available in the environment, then bacteria pass on the electrons to metabolic intermediates such as pyruvate (an internal electron acceptor) in a process called fermentation. Bacteria that ferment release acids, alcohols, and gasses. These fermentation products can be used as substrates by other microbial groups. A well-known example of these “syntrophic” associations are fermenting microorganisms that excrete carbon dioxide and hydrogen, which, in turn, can be used by methanogenic Archaea, which produce methane out of these substrates.

When electron-rich petroleum hydrocarbons are spilled into the subsurface and pool in an aquifer, the available electron acceptors are depleted by the metabolic activities of specific bacterial populations, preferentially in the order of their energy potential. For example, because the use of oxygen as a terminal electron acceptor results in the greatest energy gain for aerobic bacteria, aerobic bacteria will successfully outcompete all other microbial populations. When all oxygen is used up, and nitrate is available in the groundwater, then bacterial populations able to use nitrate as terminal electron acceptor are favored. This process will continue with Mn(IV), Fe(III), and sulfate, if they are present in the contaminated environment. When all available external electron acceptors are depleted, then the only process left is fermentation/methanogenesis.

Figure 9.1 shows the predominant TEAPS and their products that occur in a confined aquifer contaminated by petroleum hydrocarbon. The conceptual model shows sequential redox zonation from oxygen on the outside to methanic in the core of the contaminant plume (Fig. 9.1a). In practice, however, the redox zonations are not so discrete and there is often overlapping of different redox zonations (Fig. 9.1b) because biodegradation is not necessarily driven by thermodynamics alone but also by the spatial and temporal availability of electron acceptors and microbes in the impacted subsurface (Meckenstock et al. 2015). Thus, it is not uncommon to have zones of iron/manganese reduction occurring concurrently and overlapping with methanogenesis. In addition, seasonal hydrologic recharge events may shift the contaminant mass as well as deliver fresh electron acceptors to the microorganisms causing a switch from one TEAP to another (e.g., switch from methanogenesis to iron reduction and vice versa; Meckenstock et al. 2015; Teramoto et al. 2020; Beaver et al. 2021). While dissolved gasses, such as oxygen, and ions such as nitrate and sulfate easily diffuse to the bacterial membrane-bound reductases, an interesting problem arises from the use of solid electron acceptors. Some bacteria are able to reduce Fe(III) and Mn(IV) that occur in solid form as terminal electron acceptors—these bacteria can “breathe” metals.

Fig. 9.1

a Discrete redox zonations often associated with microbial oxidation of hydrocarbons based on redox ladder concept. b A new representation of the redox concept showing multiple terminal electron acceptor processes occurring in the same region of the contaminant plume (From Meckenstock et al. 2015 with permission)

Well-known and thoroughly studied iron-reducing bacteria are Geobacter and Shewanella. The iron-reducing bacteria have developed several solutions to the problem of dealing with insoluble electron acceptors. They can either (1) use diffusible compounds such as humic acids as electron shuttles, (2) be in close contact with the iron minerals, or (3) use extracellular appendages such as nanowires or an array of extracellular multiheme cytochromes to relay the electrons onto the metal ions in solid minerals (Lovley and Walker 2019). Interestingly, many methanogens are able to reduce iron, and it seems that they prefer to switch to the more energetic advantageous iron reduction as soon as they encounter minerals containing oxidized iron e.g., goethite, nontronite, illite or smectite (Prakash et al. 2019; Ferry 2020; Liu et al. 2011a, b; Zhang et al. 2012, 2013).

To put it succinctly, microorganisms mediate geochemical reactions by transferring electrons between compounds to meet their energy demands. Geophysical techniques could detect the intermediate and end products of these biogeochemical reactions (biosignatures) in the subsurface. In the next section, we will discuss different microbial-mediated physical property changes of soil and sediments and their resulting geophysical signatures.

9.1.2 By-Products of Microbial-Mediated Redox Processes Drive Geophysical Property Changes

Bacterial metabolic activities can influence the geophysical properties of the subsurface environment by changing the pore fluid chemistry through enhanced mineral dissolution, altering the composition and properties of solids (mediating precipitation of different mineral phases or changes of existing mineral phases) and influencing the overall microbial presence and their activities (microbial cells and biofilms).

9.1.2.1 Pathway 1: Pore Water Chemistry Changes

As shown in Table 9.1, several of the microbial generated metabolic products can change the pore fluid chemistry. As an example, the process of fermentation produces organic acids such as acetic acid or formic acid (Madigan et al. 2018). In addition, CO₂ is an end product of hydrocarbon mineralization. The produced CO₂ dissolves in water to produce carbonic acid. The acids dissolve in pore water, lowering the pH and enhancing the dissolution and leaching of aquifer minerals releasing ions (e.g., Ca²⁺, Mg²⁺, K⁺, Si⁴⁺) into solution (Sauck 2000; Atekwana et al. 2000; Atekwana and Atekwana 2010), causing an increase of ion concentration in the pore water. Similarly, redox reactions produce ions that directly go into solution. Dissimilatory nitrate reduction produces NH₄⁺, while Mn⁴⁺ is reduced to Mn²⁺ and Fe³⁺ is reduced to Fe²⁺. The dissolved ions and organic acids elevate the pore water electrical conductivity (specific conductance), causing a decrease in the bulk resistivity of the formation (e.g., Sauck et al. 1998; Atekwana et al. 2000; Sauck 2000; Atekwana and Slater 2009; Atekwana and Atekwana 2010 and references there in) as illustrated in Fig. 9.2. As a result, hydrocarbon impacted subsurface environments undergoing significant intrinsic bioremediation (or natural attenuation) are typically characterized by changes in pore fluid chemistry and elevated pore water conductivity (Sauck 2000), making electrical geophysical techniques (ERI, GPR, EMI, IP, or SP) ideal for their characterization and monitoring. In Sect. 9.3, we provide select case studies of geophysical signature changes driven by changes in pore fluid conductivity.

Table 9.1 Common microbial metabolic processes in the subsurface, with electron acceptors used and metabolic products that can be found in the environment. The information in this table is compiled from Mann et al. (1988), Lee et al. (2011), Madigan et al. (2018), Yu and Leadbetter (2020), Hoffmann et al. (2021) and references therein

Fig. 9.2

Spatial distribution of subsurface terminal electron acceptors and weathering products in groundwater from the FT-2 site at the Wurtsmith Air Force Base, Oscoda, Michigan, USA, a site of active biodegradation: a location of site in Michigan, pink outlines the FT-2 plume. The plume resulted from JP4 jet fuel spilled as a result of fire training exercises from 1958 to 1991. b The geochemistry data were acquired in 2007. The core of the plume (well ML-15 and 12) shows lower dissolved oxygen and nitrate resulting from the utilization of oxygen and nitrates by indigenous microorganisms. Reduction of manganese(IV) by manganese reducing bacteria and iron(III) by iron reducing bacteria results in elevated concentrations of manganese(II) and iron(II) respectively, whereas utilization of sulfate by sulfate reducing bacteria results in the sulfate depletion. The terminal electron acceptor processes are accompanied by acid production, lower pH and elevated concentrations of CO₂, accelerating the weathering and dissolution of calcium and silica, increasing the ionic content expressed as an increase in the specific conductance. Details on the geochemical analysis can be found in Che-Alota et al. (2009). Figure modified from Atekwana and Atekwana (2010) with permission

9.1.2.2 Pathway 2: Microbial-Mediated Mineral Precipitation

Bacteria are involved in the dissolution and precipitation of minerals (Hoffmann et al. 2021). Environmental conditions that are conducive for mineral precipitation include ion concentrations that are higher than their solubility and the presence of crystallization nuclei. Bacterial cell surfaces can provide both. Bacterial cell walls contain a variety of negatively charged groups, such as phosphate and carboxyl groups that are stabilized by divalent cations. These can serve as nucleation sites and high local concentrations of ions. In addition, metabolic activities of bacteria can influence the pH of the local environment, e.g., the pore water. For example, when bacteria degrade proteins or urea, ammonium ions are released, which may lead to a locally higher pH, which, in turn, can lead to the increased precipitation of minerals such as carbonates. An extensive list of minerals that have been found to be precipitated in association with bacteria is shown in the recent review by Hoffmann et al. (2021). The list includes a variety of carbonates, phosphates, silicates, sulfides, sulfates, and oxides. As one example, the product of iron reduction, Fe(II), has a much higher solubility at neutral pH than Fe(III). Dissolved Fe(II) atoms can adsorb to ferrihydrite [(Fe³⁺)₂O₃·0.5 H₂O] and induce the transformation of ferrihydrite to lepidocrocite, goethite, or magnetite (Hansel et al. 2003, 2005). Hansel et al. (2005) found that the transformation of ferrihydrite to goethite and lepidocrocite occurred at low Fe(II) concentrations (≤ 0.2 mM), while concentrations of ≥ 2 mM Fe(II) resulted in the formation of magnetite. The precipitation of magnetite can be used to infer the presence of iron reduction as a TEAP in the subsurface (Atekwana et al. 2014) which might be useful for the design of remediation programs. Nonetheless, although magnetite precipitation is typically associated with iron reducing bacteria such as Geobacter and Shewanella, there is now mounting evidence indicating that methanogens can switch their metabolism from methanogenesis to iron reduction causing the precipitation of magnetite (Beaver et al. 2021; Amiel et al. 2020; Shang et al. 2020; Sivan et al. 2016).

Magnetite is conductive and magnetic and thus the presence of magnetite can be detected by using electrical and magnetic geophysical techniques. In contrast, pyrite is conductive and not magnetic thus can only be detected by electrical geophysical techniques. Precipitated minerals like calcite are non-conductive and non-magnetic and can be detected by electrical geophysical techniques (e.g., Saneiyan et al. 2019). In essence, the ability of geophysical techniques to detect zones of microbial-mediated mineral precipitation depends on the physical property of the precipitated minerals and whether they occur in sufficient concentrations to be detected (Abdel Aal et al. 2014; Revil et al. 2015).

9.1.2.3 Pathway 3: Microbial Cells and Biofilms

Although bacteria have been studied for more than 100 years in liquid laboratory media (planktonic cells), it turns out that in nature, most bacteria are sessile and form biofilms or bioclusters (Flemming et al. 2016; Baveye 2021). Biofilms are defined as a community of microorganisms that stick to a surface in a self-produced matrix. The matrix consists of excreted biomolecules, including polysaccharides, proteins, lipids, and deoxyribonucleic acid (DNA). Biofilms survive adverse environmental conditions much better than planktonic cells. For example, biofilms are more resistant to antibiotics, can withstand drying out, and are more protected against grazing by protozoa. These biofilms, consisting of microbial cells and the matrix they produce, can clog pore spaces in sediments and aquifer matrices and therefore change the geophysical properties of the subsurface contaminated with hydrocarbons (e.g., Fig. 9.3). In both natural and engineered systems, biofilms are a key factor in clogging of sediment pore spaces and fluid flow pathways (Baveye et al. 1998). Studies have shown that biofilm development may significantly reduce the porosity (by 50–90%) and permeability (by 95–99%) of porous media (Bouwer et al. 2000; Dunsmore et al. 2004) and thus may negatively impact the efficacy of remediation programs. Therefore, determining where these biofilms are forming in the subsurface from different amendment strategies is an important consideration.

Fig. 9.3

Scanning electron micrograph image of biofilm growth in the pores of sediments with the potential to clog the pores of sediments, altering the hydraulic conductivity. Figure adapted from Sharma et al. (2021) with permission

9.2 Geophysical Methods

While geophysical techniques can be used for investigating the subsurface lithology and characterization of fractured medial to determine preferential contaminant flow paths, in this section, we provide a summary only of surface geophysical methods that are commonly used for the determination of the presence of LNAPL-contaminated sediments and for the monitoring of the progress of bioremediation. Admittedly, a number of direct push logging tools are also used but are beyond the scope of this chapter. A more detailed overview of geophysical methods can be found in most environmental and near surface geophysics textbooks (e.g., Reynolds 2011; Rubin and Hubbard 2005; Binley and Slater 2020).

9.2.1 Electrical Methods

Electrical methods measure voltages associated with electrical current flow in the subsurface to elucidate subsurface geoelectrical properties. Electrical methods can measure currents injected (or induced) into the subsurface (ERI, EMI, GPR, IP) while others like SP and Magnetotelluric (MT) can measure naturally occurring electrical and electromagnetic fields.

9.2.1.1 Electrical Resistivity

Electrical resistivity is one of the most versatile and commonly used electrical methods applied in the study of hydrocarbon-contaminated environments. The fundamental principle of the electrical resistivity method is based on the ease or difficulty of charges to flow through a material. Electric current flow in the subsurface occurs in three ways: (1) electronic conduction in materials containing free electrons such as metals, (2) electrolytic (ionic) conduction by ions in the pore waters, and (3) surface conduction (interfacial conduction) occurring in the electrical double layer at the interfaces of minerals in contact with the pore water (important for clays and disseminated metals). In the absence of metallic minerals, electrolytic conduction is by far the most common mechanism by which current flows in subsurface geologic media.

In practice, the electrical resistivity method involves the injection of direct current into the ground through electrodes connected to an artificial source of current and determining the apparent resistivity by measuring the potential at other electrodes in the vicinity of the current (Telford et al. 1990; Loke 2001). Apparent resistivity is the resistivity that would yield the measured relationship between applied current and the potential difference for a particular arrangement and spacing of electrodes of an electrically homogeneous and isotropic half-space. The measured apparent resistivity depends on the physical properties of the material and is obtained as the product of measured resistance (R) and the geometric factor (K) for a given electrode configuration (Eq. 9.1), and its unit is ohm·m.

$${\rho }_{a}=RK, R=\frac{\delta V}{I}$$

(9.1)

where ${\rho }_{a}$ is the apparent resistivity, $R$ is the resistance, $K$ is the geometric factor dependent on electrode configuration, $\delta V$ is the potential difference, and $I$ is the induced current.

There are several electrode configurations used for several application purposes, such as Wenner, Schlumberger, Wenner-Schlumberger, pole-pole, dipole–dipole arrays, or a combination of these. A detailed analysis of the advantages and disadvantages of each electrode configuration is discussed by Loke (2001). The apparent resistivity values obtained from field measurements are not “true” or actual resistivity, but the true resistivity is estimated by carrying out an inversion of the apparent resistivity using available software packages. Some commonly used inversion software are RES2DINV (Loke 2001), AGI EarthImager, and the open sourced ResIPy (Blanchy et al. 2020). In most cases the electrical resistivity is dependent on the porosity, the degree of saturation, and the conductivity of the fluid filling the pores. Thus, it is expected that the high resistivity of the LNAPLs replacing the more conductive pore waters will result in a resistive response. However, biodegradation and the production of additional ions will enhance electrolytic conductivity and will result in an expected conductive response. Nevertheless, the response is also a function of the host mineralogy/geology. For example, clays, certain remediation amendments, and salt water are very conductive and so even a relatively conductive plume may still result in a resistive response.

9.2.1.2 Ground Penetrating Radar

The GPR method is a non-destructive method that uses pulsed electromagnetic waves in the MHz-GHz range to image the subsurface. The GPR system comprises a signal generator, transmitting and receiving antennas, and a control console to display the signal generated (Reynolds 2011). The GPR antenna transmits a high frequency (typically 25 MHz to 1 GHz) electromagnetic (EM) wave that propagates into the subsurface. When the propagating EM wave encounters a boundary or interface with contrasting electromagnetic properties [magnetic permeability ($\mu$), dielectric permittivity $(\varepsilon$), or electrical conductivity ($\sigma$)], some of the transmitted energy is reflected to the surface. Since in most materials the magnetic permeability does not vary much, the dielectric permittivity is the property that determines the EM wave velocity in most materials (Eq. 9.2).

$$V=1/\sqrt{\mu \varepsilon }$$

(9.2)

Commonly the dielectric behavior is characterized in terms of the relative dielectric permittivity (also known as the dielectric constant):

$$\kappa =\varepsilon /{\varepsilon }_{0}$$

(9.3)

where $\varepsilon$ ₀ is the permittivity of a vacuum (8.8542 × 10^–12 F/m). Neglecting the effects of magnetic permeability ($\mu$) since near surface sediments and soils are typically non-magnetic, the EM wave velocity can be rewritten as:

$$V=c/\sqrt{\kappa }$$

(9.4)

where c is the velocity of light in a vacuum (0.3 m/ns). Water has a significantly higher dielectric constant (κ = 80) than soils and sediments (κ = 4–10) and LNAPLs (κ ~ 2), hence GPR techniques are commonly used for providing information on water content (saturation) (Davis and Annan 1989).

The reflected waves across interfaces and the two-way travel times are recorded by a receiver antenna. The amplitude of the reflection depends on the magnitude of the contrast of electromagnetic properties across the boundary. The attenuation factor ($\alpha$) and the depth of penetration is dependent on the electrical conductivity ($\sigma$), magnetic permeability ($\mu$), and dielectric permittivity ($\varepsilon$) of the media through which the signal is propagating (Reynolds 2011) as shown in Eq. (9.5).

$$\alpha \approx \frac{\sigma }{2}\sqrt{\frac{\mu }{\varepsilon }}$$

(9.5)

From Eq. (9.5), it is evident that attenuation is related to conductivity; the higher the conductivity is, the greater the attenuation and the shallower the depth of penetration will be (Pettersson and Nobes 2003). GPR signals are greatly attenuated in highly conductive materials such as clayey soils and sediments and by high salinity pore waters, limiting their penetration and use in such environments. At hydrocarbon-contaminated sites where biodegradation increases the pore water conductivity, the amplitude of the reflected energy is attenuated resulting in low signal amplitudes (muted reflections, or shadow zones) coincident with the zone of contamination (Sauck et al. 1998; Sauck 2000).

9.2.1.3 Electromagnetic Induction

EM methods are commonly used at hydrocarbon-contaminated sites to map shallow LNAPL contamination, as well as locate underground utilities such as buried pipes including underground storage tanks (e.g., Atekwana et al. 2002). The basic principles of shallow EM techniques can be found in McNeill (1980, 1991). Briefly, the EM method uses the response of the ground to the propagation of incident alternating electromagnetic waves that are made up of two orthogonal components consisting of an electric field (E) and a magnetic field (H) in a plane perpendicular to the direction of the EM wave propagation. A transmitter coil is used to generate the primary electromagnetic field, which propagates above and below ground. When a subsurface conductor is encountered, the primary magnetic field induces eddy currents in the conductor. The eddy currents in turn induce a secondary magnetic field which is measured by the receiver coil but is delayed in phase. The receiver coil measures a resultant field which is made up of the primary (from the transmitter coil) and secondary magnetic field generated by the subsurface conductor. The phase lag between the primary and secondary fields is used to make useful deductions about the subsurface conductor. In general, the depth of penetration of the EM wave is a function of the frequency and conductivity. The skin depth Z_s is defined as the depth at which the amplitude of the plane wave is reduced to 1/e or 37% of its initial amplitude. This depth of penetration of an EM wave is given by:

$${Z}_{s}=503.8 \sqrt{f\sigma }$$

(9.6)

where f is the frequency of the EM wave in Hz and σ is the conductivity (Reynolds 2011). The depth of penetration of EM systems is also a function of the distance between the receiver and transmitter coils as well as the orientation of the coils.

In ground conductivity meters such as the EM-31 manufactured by Geonics Ltd, which is commonly used at hydrocarbon-contaminated sites, the receiver coil measures both the magnitude and phase of the secondary field and provides two readings: the quadrature component, which is related linearly to the apparent conductivity is measured in units of millisiemens per meter (mS/m) and the in-phase component, which is a measure of the metal content in parts per thousands (ppt) provides useful information for determining if the source of the anomaly is a metal (e.g., buried metallic pipe). The maximum depth of penetration for the EM-31 signals is ~6 m.

9.2.1.4 Induced Polarization

Although direct current (DC) resistivity methods are most used to characterize LNAPL sites, they constitute a bulk measurement, responsive to both electrolyte and solid–fluid interface (surface) chemistry, and therefore are unable to differentiate between the relative contributions of electrolytic versus interface conductivity or discriminate between clay-rich sediments and more saline pore water. The IP technique, which can be measured either in the time domain or frequency domain, is an extension of the DC-resistivity method. It is more sensitive to changes in the electrochemistry of the pore water–mineral interface (e.g., Lesmes and Frye 2001) and allows for the complex electrical properties (electromigration and polarization) to be measured (e.g., Binley and Kemna 2005; Revil et al. 2012). Because biodegradation potentially modifies the mineral surface properties and because microorganisms typically are attached to the mineral grains, the induced polarization technique, particularly the spectral induced polarization (SIP, a multifrequency measurement) has been documented as being sensitive to the presence of NAPLs and also suitable for investigating the effects of bio-physicochemical changes of electrical properties in hydrocarbon-contaminated sediments (Abdel Aal et al. 2004, 2006; Orozco et al. 2012; Johansson et at. 2015). In the SIP technique, the impedance magnitude |σ| and the phase shift φ (between a measured sinusoidal voltage and an induced sinusoidal current) is measured over a range of frequencies, typically between 100 MHz and 1000 Hz.

From the magnitude and phase measurements, the real (σ′ = |σ| cos φ) and imaginary (σ″ = |σ| sin φ) parts of the sample complex conductivity (σ*) are calculated:

$$\sigma^{*} (\omega ) = \sigma^{^{\prime}} (\omega ) + i\sigma^{^{\prime\prime}} (\omega )$$

(9.7)

where, the in-phase (real; σ′) conductivity component represents electromigration in the subsurface and is sensitive to changes in fluid chemistry, whereas the out-of-phase (imaginary, σ″) conductivity represents the charge polarization which at low frequencies (<1000 Hz) results primarily from the polarization of ions in the electrical double layer (EDL) at the mineral–fluid interface and i is the square root of √ − 1 (Lesmes and Frye 2001). Conduction and polarization at the fluid–grain interface are a function of surface area, pore size geometry, surface charge density, and surface ionic mobility (Lesmes and Frye 2001). The real conductivity is what is typically measured in electrical resistivity imaging. Several laboratory experiments have clearly established that σ″ is more sensitive to microbial activity and presence of biofilms than σ′ (Abdel Aal et al. 2004; Davis et al. 2006; Zhang et al. 2014; Mellage et al. 2018, 2019; Kimak et al. 2019).

Most modern electrical resistivity instruments are equipped to make time domain IP (TDIP) measurements during ERI surveys. TDIP is acquired simultaneously with resistivity measurements by measuring the transient decay of the voltage after the current is shut-off, typically in the form of the integral of the decay curve over a defined time window.

For TDIP applications, the parameters measured are the apparent chargeability (${m}_{a}$) calculated as:

$${m}_{a}=\frac{{V}_{s}}{{V}_{p}}$$

(9.8)

where, ${m}_{a}$ is the apparent chargeability (with unit mV/V) at time t, ${V}_{s}$ is the secondary voltage measured at time t after the current is off, ${V}_{p}$ is the primary voltage measured when the current is on. It should be noted that ${V}_{s}$ is only significant in polarizable subsurface environments. At low frequencies (below 10 Hz), the measured parameters in the FDIP and TDIP are proportional and related by the equation:

$${m}_{a}=-\kappa \varphi$$

(9.9)

where, $\kappa$ is the constant of proportionality, which can be experimentally derived. Because FDIP measurements are more challenging to acquire in the field, TDIP chargeability values acquired during electrical resistivity surveys can be transformed to phase angle measurements using Eq. (9.9) (Binley and Kemna 2005).

9.2.1.5 Self Potential

The SP method involves the measurement of differences in natural electric potentials developed in the Earth. The potentials measured can range from a few millivolts (mV) to greater than 1 V (Reynolds 2011). The sign of the potential (negative or positive) is diagnostic of the source generating the potential. SP anomalies can be generated in the Earth through the following processes:

• (1) Electrokinetic (electrofiltration, streaming) potentials result from electrolytes flowing through a capillary tube or a porous medium. These types of potentials are transient and are typically generated during groundwater flow associated with recharge events, dam seepage, groundwater pumping, or discharge zones.

• (2) Electrochemical (diffusion or liquid junction) potentials result from local differences in the mobilities of electrolytes (anions and cations) due to differences in concentrations. These potentials are small (in the tens of mV range) and can occur at contaminated sites where differences in the concentrations of ions within and outside the plume exist. The electrochemical potentials can also be generated at shale–sand interfaces and are the major source of SP anomalies measured in well logging.

• (3) Thermoelectric potential results from temperature gradients such as in geothermal areas. In addition, microbial activity in hydrocarbon-contaminated environments may result in thermal anomalies (e.g., Warren and Bekins 2018) that could result in temperature gradients between contaminated and background regions generating thermoelectric potentials. This type of potential is also transient and exists as long as the thermal differences exist.

• (4) Mineral potentials are typically very large (100 s mV), associated with massive sulfide ore bodies, and used in the exploration of these deposits. While their origin is not well understood, they are hypothesized to be caused by the existence of geobatteries (Sato and Mooney 1960), where an ore body straddles the water table serving as an electronic conductor for electron transport from anodic reactions below the water table to cathodic reactions occurring above the water table. Naudet et al. (2004) and Revil et al. (2010) have documented the existence of large SP anomalies at contaminated sites resulting from biogeobatteries associated with redox reactions. Here, the electronic conductors are biofilms, conductive bacterial appendages (nanowires), or bio-metallic minerals within the water table fluctuation zone that move electrons from reducing environments below the water table to oxidizing environments above the water table.

At hydrocarbon-contaminated sites, the cause of SP may be related to electrokinetic potentials, electrochemical potentials, thermoelectric potential or mineral potentials or a combination of these processes.

Self-potential measurements are easy to make. The equipment required for measuring SP anomalies consists of a pair of non-polarizable electrodes (e.g., Cu/CuSO₄ porous pots) connected by insulated cable to a high input impedance (~100 MΩ) voltmeter, which is used to read the potential with a 0.1 or 1 mV resolution. Field surveys are commonly done in two ways: gradient surveys and fixed base station mode. In gradient surveys, two electrodes (with constant spacing between them—typically 5–10 m) are moved successively along a survey line and the potential difference between the two electrodes is divided by the spacing between the electrodes expressed in mV/m. Electrodes are moved along the line in leapfrog fashion. In the fixed electrode configuration, one electrode is fixed at a location outside the target zone referred to as the base station and the potential difference between the base station and roving electrode is measured. SP measurements are susceptible to cultural noise associated with stray currents such as those from electric trains, or geological noise from changing soil conditions (wet to dry, forested to grasslands, elevation difference) and telluric currents; hence care must be taken in the processing to identify and eliminate noise.

This unsophisticated methodology is also reflected in the results, which are often interpreted qualitatively and provide general information about anomaly shape and amplitude. However, more advanced processing and inversion can be used to retrieve the current source densities and quantitative information obtained on the redox potentials at the site (e.g., Naudet et al. 2004; Revil et al. 2010; Giampaolo et al. 2014; Abbas et al. 2017, 2018).

9.2.2 Magnetic Method

Magnetic methods obtain information related to the intensity, direction, gradient of the magnetic field of the Earth. In the context of application to hydrocarbon contamination, magnetic methods exploit variations in the magnetic mineralogy of materials. Some common magnetic minerals which may cause variation in the measured magnetic properties include magnetite, maghemite, hematite, pyrrhotite, and greigite. One of the magnetic methods most applied in investigating hydrocarbon contamination is magnetic susceptibility.

Magnetic susceptibility is defined as the degree to which a sediment or mineral can be magnetized when a magnetic field is applied. Magnetic susceptibility ($\chi$) is the ratio of applied magnetization ($M$) to the applied magnetic field ($H$):

$$M=\chi H$$

(9.10)

where, $M$ and $H$ have the same units (A/m), and $\chi$ is the volume-specific dimensionless magnetic susceptibility.

Magnetic susceptibility measurements can be made on cores retrieved from hydrocarbon-contaminated sediments or by downhole measurements at field sites. In downhole measurements, a magnetic susceptibility probe is lowered down a borehole that is free of any metallic materials (including well casings). The role of magnetite cannot be overstated for hydrocarbon-contaminated sites, as even minor precipitates of magnetite can produce large magnetic susceptibility measurements. This enhanced magnetite production has been observed at many hydrocarbon-contaminated sites (Mewafy et al. 2011; Rijal et al. 2010, 2012; Atekwana et al. 2014; Lund et al. 2017) and is a key feature in the application of magnetic methods in hydrocarbon-contamination monitoring.

9.3 Geophysical Applications and Case Studies

A variety of geophysical techniques have been successfully applied at hydrocarbon-contaminated sites to infer the presence of fresh and biodegraded hydrocarbons, as well as to monitor their remediation. Although numerous laboratory experiments have been conducted to determine the efficacy of geophysical techniques to detect both fresh and biodegraded hydrocarbon-contaminated media (Atekwana and Atekwana 2010), in this chapter, we focus on field characterization and provide example case studies from different environments (fresh to saline aquifers), geographic and climatic regions (desert and cold). We have presented, in Sect. 9.1 above, the need to understand the processes occurring at hydrocarbon-contaminated sites, as this will determine the appropriate geophysical techniques to use.

9.3.1 Geophysical Signatures of Changes in Pore Fluid Conductivity

9.3.1.1 Conductive Response in Contaminated Freshwater Aquifers

An examination of the published peer review literature indicates that the conductive plume model (Sauck et al. 1998; Atekwana et al. 2000) is the dominant response observed at the majority of aged contaminated sites irrespective of geographic location. Hence, electrical techniques, which are sensitive to changes in pore fluid conductivity are best for characterizing aged hydrocarbon-contaminated sites (e.g., Werkema et al. 2003). The GPR and ERT are the two most commonly applied techniques due to their relative ease of use even for the non-practitioner. We discuss two case studies where electrical resistivity and GPR were used to investigate aged LNAPL plumes.

The first case study is a site at the decommissioned Wurtsmith Air Force Base in Oscoda, Michigan, USA, near the shores of Lake Huron. This site was the location of the seminal work of Sauck et al. (1998) that provided strong evidence linking the measured geophysical signatures to microbial degradation of the hydrocarbons and resulted in the development of the conductive plume model by Sauck (2000). Hydrocarbon contamination resulted from more than three decades of bi-weekly fire training exercises with JP-4 as the dominant contaminant resulting in what is known as the FT-02 plume. By the 1990s, the free product plume was 0.3 m thick. Much of the hydrocarbon contamination was restricted to the upper parts of the saturated zone and within the capillary fringe zone (McGuire et al. 2000; Skubal et al. 2001). The contaminant plume at FT-02 was approximately 75 m wide and extended 30 m upgradient and about 450 m downgradient to the southeast of the source area (Che-Alota et al. 2009; Skubal et al. 2001) (Fig. 9.2a). Chemical analyses of the groundwater showed elevated benzene, toluene, ethylbenzene, and xylenes (BTEX) concentrations in addition to elevated groundwater conductivities. The site geology consists mostly of clean well-sorted fine to medium sands coarsening downward to gravel of eolian origin approximately 20 m thick, underlain by a confining unit that is a brown to gray lacustrine silty clay unit 6–30 m thick (Bermejo et al. 1997; Sauck et al. 1998) with the water table at ~3–5 m below ground surface (bgs). The geophysical results shown in Fig. 9.4 were acquired in 2003 and presented in Che-Alota et al. (2009) and Atekwana and Slater (2009).

Fig. 9.4

a Electrical resistivity and b GPR image of the FT-02 plume at the decommissioned Wurtsmith Air Force Base in Oscoda Michigan, USA (modified from Atekwana and Slater 2009 with permission). The zone of attenuated GPR reflections is coincident with a region of low resistivity of the resistivity profiles that is coincident with the location of the plume. Profile location is shown on Fig. 9.2a as B–B′

The electrical resistivity survey was conducted using a Syscal R2 resistivity meter manufactured by IRIS with 72 electrodes using a dipole–dipole array and a 10 m electrode spacing. The apparent resistivity measurements were inverted using RES2DINV (Loke and Barker 1996) to retrieve the true resistivities at the site. The GPR survey was acquired using a Geophysical Survey System Inc. (GSSI) Sir 10 A + system with 100 MHz bistatic antennae recording for a total of 400 ns. A fixed 1.4 m separation between RX-TX pairs was used and gains were automatically set at the beginning of the line and the data were processed using the RADAN software. Extensive geochemistry was acquired over the site and the data are presented in Fig. 9.2. The zone of contamination is characterized by a zone of low resistivity (< 150 Ω·m) and attenuated GPR reflections (see arrows in Fig. 9.4). The geochemistry of groundwater indicates that biodegradation is active at the site. The biodegradation by-products, e.g., Fe²⁺, Mn²⁺ and weathering products (e.g., silica and calcium) have significantly perturbed the groundwater geochemistry at the site resulting in elevated ion chemistry within the plume (~4 × background) that is detected by the geophysical measurements. The attenuated GPR reflections result from the elevated conductivity within the plume.

The second case study we present is a decommissioned refinery near Paris, France (Abbas et al. 2018). Operations at the refinery started in the 1920s. Here the contamination consists of LNAPLs and other hydrocarbon products. The free phase plume had a thickness range of 0.1–1 m. The geology consists of mostly fine to medium-grained sands coarsening to gravel with a thickness ranging between 8 and 25 m (Abbas et al. 2018). This layer hosts an alluvial aquifer at a depth ranging from 4 to 11 m. A chalk aquifer underlies the alluvial aquifer at a depth ranging from 11 to 25 m. Extensive geochemical data exist for the site and were reported in Abbas et al. (2018). The electrical resistivity data were acquired with a Syscal Pro system using a Wenner and dipole–dipole arrays with a 2 m electrode spacing. The acquired apparent resistivity data were inverted using the RES2DINV (Loke and Barker 1996). The GPR data were acquired using a Mala Geoscience’s GPR system with a 250 MHz antenna. The groundwater geochemistry indicates biodegradation was active at the site with low TEA concentrations (e.g., oxygen, nitrate, sulfate) and elevated total dissolved solids.

The electrical resistivity profile (Fig. 9.5a) and GPR profile (Fig. 9.5b) over the clean areas (0–75 m X position) showed high electrical resistivity (>150 Ω·m) that markedly decreased and strong GPR reflections that were attenuated within the contaminated zone (X position 75 m to end of profile). The zone of attenuation started at ~2 m beneath the surface. Note that the water table was approximately 8 m beneath the surface, hence this attenuation began in the unsaturated zone extending into the saturated zone. The resistivity profile showed a region of low resistivity (< 50 Ω·m) coincident with the region of attenuated GPR reflections consistent with a conductive response related to elevated ion concentration.

Fig. 9.5

a Electrical resistivity, b GPR image from a decommissioned refinery near Paris, France. The zone of attenuated GPR reflections is coincident with a region of low resistivity which is coincident with the location of the plume (modified from Abbas et al. 2018 with permission)

Both case studies presented suggest that in freshwater aquifers where biodegradation of the LNAPL is active, the zone of LNAPL contamination can be characterized by low resistivity and attenuated GPR reflections. These responses have been reported in many freshwater aquifers globally and across many climatic zones including dry, desert environments in Iraq (Al-Menshed and Thabit 2018), tropical environments such as in Brazil (e.g., Lago et al. 2009; Moreira et al. 2019, 2021), temperate environments such as in Serbia (Burazer and Burazer 2017), France (Blondel et al. 2014), and warm temperate–subtropical environments such as Jiangsu Province, China (e.g., Shao et al. 2021).

9.3.1.2 Characterizing LNAPL Sites Using Spectral Induced Polarization (SIP) and Time Domain Induced Polarization (TDIP)

While the electrical resistivity imaging provides information on bulk electrical property changes that are largely controlled by the electrolytic conductivity, it is unable to discriminate between fluid and lithologic effects. In contrast, SIP allows for the measurement of additional parameters such as the real (σ′) and imaginary (σ″) components of the complex conductivity (σ*), which can provide information on physicochemical processes occurring at the mineral grain–pore fluid interface. As such it can provide additional information that can allow for the discrimination of fluid conductivity and lithology (Slater and Lesmes 2002), as well as discerning microbial effects (Abdel Aal et al. 2004, 2006; Davis et al. 2006; Atekwana and Slater 2009; Orozco et al. 2012, 2021; Mewafy et al. 2013; Johansson et al. 2015). A large body of literature exists examining SIP response to soil-hydrocarbon mixtures in laboratory settings (Vanhala 1997; Schmutz et al. 2010; Abdel Aal et al. 2014; Deng et al. 2018 and references there in). These studies all suggest that SIP is excellent for detecting the presence of hydrocarbons in porous media. Despite this large advantage of SIP over ERT, only a limited number of studies exist of SIP measurements of hydrocarbon contamination in field settings (e.g., Flores-Orozco et al. 2012; Blondel et al. 2014; Maurya et al. 2018). This is largely due to the limited availability of field SIP instrumentation and problems of noise from electromagnetic coupling at higher frequencies.

In Fig. 9.6, we present a case study by Orozco et al. (2021) of SIP measurements at a kerosene-contaminated site downgradient of a military base in Decimomannu, Italy. The subsurface of the site is characterized by a backfill layer with a thickness ranging between 0.2 and 1 m and underlain by recent alluvium composed mostly of gravels and sands which extend to a maximum depth between 4 and 6 m. The sand and gravel layers are underlain by “Hazelnut clay” lenses, composed of sandy-gravelly clays of hazelnut color with a thickness of 1–1.5 m. Total Petroleum Hydrocarbon (TPH) concentrations in the subsurface ranged from 100 to 18,000 µg/L. Fluid electrical conductivities were highest at the clean locations (~1860 µS/cm) compared to contaminated locations (~1200 µS/cm). The results presented include the real, imaginary, and phase response at 15 Hz (detailed results are presented in Orozco et al. 2021). The results showed that except for the phase (φ) response, the real (σ′) and imaginary (σ″) components of the complex conductivity for the contaminated sediments (Fig. 9.6b–d) were higher than those in clean sediments (Fig. 9.6a) and higher in the dissolved phase (Fig. 9.6c) than in the free phase (Fig. 9.6b). This is clearly demonstrated in the correlation plot of Fig. 9.6d. The authors explain the increase in real and imaginary conductivity of the contaminated locations compared to the clean location as resulting from: (1) the disconnection of the electrical double layer as a result of the immiscible oils in the pores; and (2) the increase of ions from the bulk electrolyte into the electrical double layer associated with the non-miscible oils in the pore water.

Fig. 9.6

Modified from Orozco et al. (2021) with permission

Spectral induced polarization (SIP) results from a kerosene impacted site near Decimomannu, Italy. The results shown are the 15 Hz data of the real (σ′), imaginary (σ″), and phase (φ) from a clean, b free phase, c dissolved phase, and d correlation of the real (σ′) and imaginary (σ″) component of the complex conductivity from pixel values extracted from the SIP data at different depths: between 1 and 2 m (unsaturated zone), 2.80 and 3.80 m (around the water table), and 4.5 and 5.5 m (saturated zone). The SIP parameters are extracted from the clean sediments, free-phase LNAPL, and dissolved phase plume. The black line is the linear increase related to the model σ′ = 100 σ″, which is used to indicate regions dominated by electrolytic conduction (below line) and surface conduction (along the line). The horizontal line at 3.3 m depth indicates the position of the groundwater table at the time of the survey. BH1–BH7 are lithologic boreholes superimposed on the SIP profiles, with the boxes indicating: the backfill materials on the top (white), the recent alluvial (no color), the Hazelnut clays (gray), and the ancient alluvial sediments (no color, at the bottom).

Although the above study is unable to use SIP to estimate the volume of TPH in the soil, a study by Deng et al. (2018) documents in a sandbox experiment that SIP can map an oil plume and estimate the volumetric oil content with an efficacy rate of 70–80% with the efficacy rate higher for the imaginary conductivity component.

In the second example, we describe the use of TDIP measurements as a surrogate for FDIP measurements. Here TDIP measurements were acquired during ERI surveys of the FT-02 plume in 2007 (the location of the profile A-A’ is shown on Fig. 9.2) at the Wurtsmith Air Force Base, Oscoda MI, described in Sect. 9.3.1.1. The TDIP survey was conducted using an IRIS Syscal Pro with an axial dipole–dipole array with 3 m electrode spacing. The TDIP measurements were transformed to produce phase angle, real, imaginary and responses following Binley and Kemna (2005). The results from the IP inversion are presented on Fig. 9.7. Boundaries of the plume as determined by geochemical data are marked on the figures by vertical black lines. The phase angle data (Fig. 9.7a) show a decrease in the phase angle (values range between −1 and −2.5 mRads) within a region coincident with the plume (approximately 35–95 m horizontal distance) compared to regions outside the plume (−4 to −5 mRads). At ~40 m there is an elliptical zone of anomalously high phase angles (6 to −7 mRads) that is probably due to noise from buried infrastructure. The data showed that a zone of higher real and imaginary conductivity occurring below the water table (5 m) is coincident with the plume (Figs. 9.7b, c). Outside of the plume boundary, there is no significant variation in both the real and imaginary conductivity values below the water table. We interpret the higher real and imaginary conductivities within the zone of contamination as related to increase in ion concentrations resulting from enhanced mineral weathering due to acids produced during biodegradation. Studies by Che-Alota et al. (2009) at this field site indicated that when compared to the 1996 and 2003 data, the bulk electrical resistivity within the contaminated zone had reverted to near background conditions in 2007 due to contaminant mass removal from a soil vapor extraction system installed in 2001. However, we note that the TDIP profile still showed the effects of biodegradation consistent with groundwater geochemical data (e.g., elevated CO₂, Fe²⁺, Mn²⁺, and decreased sulfate and nitrate; Fig. 9.2b), perhaps related to the higher sensitivity of IP. This may suggest that IP may be more diagnostic to microbial processes when compared to the DC resistivity.

Fig. 9.7

a Phase angle, b Real conductivity, and c Imaginary conductivity obtained from transformation of time domain induced polarization measurement at the FT-02plume site, Wurtsmith Air Force Base, Oscoda, Michigan, USA. Plume boundary is marked by vertical black lines. Location of profile is shown as A–A′ on Fig. 9.2a. Water table is marked by a dashed line and inverted triangle

Other TDIP studies from hydrocarbon-contaminated sites have shown mixed results. For example, Blondel et al. (2014) reported a clear decrease in bulk electrical resistivity related to biodegradation of an oil spill site but no measurable changes in the normalized changeability or quadrature conductivity. On the other hand, in other studies, Deceuster and Kaufmann (2012) and Abbas et al. (2018) reported both low bulk electrical resistivity anomalies and high-chargeability anomalies coincident with the contaminated aquifer. The above discussion points to the fact that, sometimes the electrical response can be ambiguous. It is therefore important to constrain geophysical interpretations with groundwater geochemistry data.

9.3.2 Resistive Response in Saline Aquifers

Although many aged hydrocarbon-contaminated sites show a conductive electrical resistivity response associated with biodegradation, not all sites are conductive. We provide examples from a coastal environment where a resistive response is observed. Figure 9.8a shows the results from an electrical resistivity survey conducted at a former perfumery plant in Jinghai, southwest of Tianjin, China, with a benzene and ethylbenzene spill history. Details of this study are presented in Xia et al. (2021). The site geology consists of a ~1.5 m thick top layer of backfill underlain by ~1 m of silt, clay and silty clay layers. Underlying this, at a depth of 13–16.5 m, is a dense silt layer forming an impermeable boundary. Contaminant concentrations obtained from groundwater samples showed their highest concentrations in the range of 2592 μg/l for benzene and 92,800 μg/l for ethylbenzene. Soil sampling showed that most of the contaminants were hosted in the silt and clay layers. The water table is ~4 m bgs and fluctuates annually within 1.9–4.1 m bgs. The groundwater has high salinity in the range of 4–10 g/L. The electrical resistivity data were acquired using a combination of a gradient and Wenner arrays with an electrode spacing of 2 m. The data were inverted using RES2DINV to obtain the resistivity structure of the subsurface (Loke and Barker 1996). The locations of borings with contaminant concentration are superimposed on the resistivity profile. The resistivity profile shows two layers, an upper resistive layer that is 5 m thick with ~12–20 Ω·m resistivity and a lower conductive layer with resistivity values <6 Ω·m. The highest resistivity values (>30 Ω·m) in the shallow subsurface (upper 2.5 m) are associated with high concentrations of benzene and ethylbenzene.

Fig. 9.8

a Geophysical signatures of LNAPL contamination in saline aquifers showing the electrical resistivity tomography profile from a former perfumery plant in Jinghai, southwest of Tianjin, China. T95, DT36, and T63 are soil borings with benzene and ethylbenzene concentrations superimposed on the resistivity profile (modified from Xia et al. 2021 with permission). High concentrations of benzene and ethylbenzene are coincident with regions of high resistivity and the region of high resistivity is associated with the oil-contaminated zone. b Electrical resistivity profile from Grande Terre Barrier Island, Louisiana, USA acquired during the 2010 BP Deep Horizon Oil spill in the Gulf of Mexico (Ross 2013). Note that the region of high resistivity is associated with the oil-contaminated zone

Despite the long history of contamination at this site, the LNAPL contamination showed a resistive response. This response may be the result of lack of significant degradation or the fact that the spill is hosted in a more conductive host lithology (silts and clays) as well as conductive groundwater. Hence any conductive signature arising from biodegradation might be masked by the conductive clays or saline groundwater.

The second example we present is from the Grande Terre Barrier Island off the southeastern coast of Louisiana, USA (Fig. 9.8b). Details of the study can be found in Heenan et al. (2015). The contamination resulted from the BP Deep Horizon spill which occurred on April 20, 2010, spilling 4.1 M barrels of oil in the Gulf of Mexico. Although dispersants were used to disperse the oil, the spill resulted in the contamination of coastal communities and barrier islands. The electrical resistivity data were acquired using a dipole–dipole array with 0.5 m spacing approximately 4 months after the spill. The subsurface geology consisted of fine to medium-grained sands saturated with salt water. The resistivity profile presented in Fig. 9.8b shows that the resistivity is generally very low (<2.5 Ω·m) due to the water salinity. However, a layer of relatively higher resistivity (~1.5–2.3 Ω·m) occurring at shallow depths (~1–2 m) is spatially correlated to the impacted layer based on soil borings and thickens toward the shoreline. In this example, the high resistivity response resulted from a fresh spill which is resistive and hosted in an aquifer with very conductive salt water. It is important to point out that continued monitoring of the site over an 18-month period documented the attenuation of the resistive anomaly as a result of biodegradation (Heenan et al. 2015).

Although we have presented only two examples of LNAPL spills and biodegradation in saline aquifers in coastal regions, another good example is a spill in the Niger Delta area of Nigeria. A resistive response was the dominant response observed in the electrical resistivity profiles (Raji et al. 2018; Uchegbulam and Ayolabi 2014). The resistive response probably resulted from the fact that this is an area of active oil exploration with repeated spills from pipeline breakage. In addition to the resistive response observed in saline aquifers, a resistive response can also be obtained when the spills are hosted in conductive lithologies such as clays and shales as demonstrated by a study in Argentina by Osella et al. (2002).

9.3.3 Example from Cold, Permafrost Environments

Pettersson and Nobes (2003) provide an example of the use of geophysical techniques (EM-31 and GPR) to investigate hydrocarbon contamination (mostly JP4 and JP5) at Scott Base, located on Ross Island in the Ross Sea region in Antarctica. Scott Base has been occupied since 1957 and anthropogenic activities have resulted in hydrocarbon contamination at the base. Figure 9.9 shows the pre-melt EM survey results from Scott Base presented as apparent conductivity. The red colors represent high conductivities and blue colors represent low conductivities.

Fig. 9.9

(Modified from Pettersson and Nobes 2003 with permission)

Electromagnetic signatures of LNAPL contamination in a cold, permafrost region. 1–4 represent locations of low electrical conductivity coincident with regions of LNAPL contamination

Areas identified and labeled as 1, 2, 3, and 4 represent regions of low conductivity (high resistivity) ranging from 1 to 3 mS/m which are interpreted to be the result of the hydrocarbon contamination. Hydrocarbon contamination was confirmed by dug pits that showed oil slicks on water. Laboratory analysis confirmed the presence of hydrocarbons with TPH concentrations ranging from 1860 to 5560 mg/kg. Although studies from other permafrost regions have documented the potential for hydrocarbon degradation in subzero temperatures (e.g., Rike et al. 2003; Børresen et al. 2003), albeit at a slower rate due to low temperatures and low nutrients, at Scott Base, the LNAPL contamination results in a resistive response. It is possible that the high resistivity (low conductivity) response at Scott Base may be due to the occurrence of repeated fresh spills, the lack of extensive biodegradation or the fact that the region is underlain by soils with high salt concentration due to the proximity of the site to the sea (Pettersson and Nobes 2003).

9.3.4 Geophysical Signatures of Microbial-Mediated Mineral Precipitation

We discussed in Sect. 9.1.2 how microorganisms can change the pore water chemistry resulting in the precipitation of different mineral phases as documented in Table 9.1. Different minerals can be precipitated depending on the TEAPs and can therefore be used at hydrocarbon-contaminated sites to delineate zones of contamination, as well as infer the TEAPs at the site. The geophysical technique used to detect zones of bio-mediated mineral precipitation depends on the physical property of the biomineral (Atekwana and Adel Aal 2015). For example, the precipitation of magnetite resulting from the activity of iron-reducing bacteria can be used to infer zones where iron reduction is occurring. Magnetite is both magnetic and conductive, requiring the use of either electrical or magnetic techniques for its detection (e.g., Mewafy et al. 2013; Atekwana et al. 2014). In fact, laboratory investigations by Porsch et al. (2010) suggest that magnetic properties such as magnetic susceptibility (MS) can be used to assist in the delineation of hydrocarbon contamination in the environment. We present, in Fig. 9.9, examples of the use of MS to delineate zones of enhanced MS at three different hydrocarbon-contaminated sites.

Figure 9.10a shows the magnetic susceptibility measurements from cores retrieved from an abandoned refinery site in Carson City, Michigan, USA. The refinery at the site was in operation for more than 60 years and was decommissioned in the 1990s. Historical releases from storage tanks and pipelines resulted in the contamination of the subsurface aquifer. This site has been the focus of several geophysical (Atekwana et al. 2000, 2004; Werkema et al. 2003; Abdel Aal et al. 2006), microbial (Allen et al. 2007), and geochemical (Atekwana et al. 2004) studies. These studies confirmed active intrinsic bioremediation at the site. In situ resistivity measurements down boreholes identified a zone of enhanced conductivity coincident with the zone of contamination and enhanced microbial activity. This conductivity enhancement was also coincident with the water table fluctuation zone (WTFZ). The magnetic susceptibility values displayed in Fig. 9.10a were obtained from cores retrieved from the site in 2007 and measured in the laboratory using a benchtop MS meter. The results showed a decrease in MS from the surface down to an elevation of 226 m. This elevation is also coincident with the top of the hydrocarbon smear zone formed as a result of seasonal water table fluctuations. At 225.5 m, the MS increased with variable excursions to an elevation of 224 m below which it decreased to the end of the measured core. This zone of enhanced MS was coincident with the zone of contamination and the zone of enhanced conductivity documented by Werkema et al. (2003). In fact, Werkema et al. (2003) suggested that the pore water conductivity in this zone was around five times the conductivity at the uncontaminated location. We can infer that the MS was the result of magnetite precipitation resulting from microbial iron reduction within the WTFZ.

Fig. 9.10

Magnetic susceptibility measured at three different hydrocarbon-contaminated sites. a Carson City, Michigan, USA; b Montana, USA; c Bemidji, Minnesota, USA. All three sites show positive excursions in the magnetic susceptibility within the water table fluctuation zone. c Adapted from Lund et al. (2017) with permission

Figure 9.10b shows in situ downhole MS measurements from a decommissioned refinery site in, Montana, USA. The primary contaminant was crude oil. The depth to groundwater was approximately 3.6–4 m below ground level with groundwater specific conductance at ~15,000 µS/cm making the aquifer at this site very saline. MS logging involves lowering a MS probe/sensor into the borehole while making MS measurements. The MS logging tool can be operated in uncased, or PVC cased wells that have a diameter of 2″ (~50 mm) or bigger. The results showed stable MS values from the surface down to a depth of 3 m at which point significant excursions occur reaching a value of 150 SI × 10^–4. This zone of enhanced MS was ~1.5 m thick and extended to a depth of 5 m. Below 5 m the values stabilized at ~100 SI × 10^–4. Geochemical data at the site suggested biodegradation was occurring at the site and the plume was methanic.

9.3.5 Geophysical Investigations at Bemidji, Minnesota, USA

Lund et al. (2017) and Atekwana et al. (2014) document similar MS responses from a site near Bemidji, Minnesota, USA, where a crude oil pipeline ruptured, releasing 1,700,000 L of crude oil into the environment. After initial clean up, at least 400,000 L of hydrocarbon remained in the subsurface and the site has been used as a natural laboratory for investigating natural attenuation processes associated with hydrocarbon contamination. The site is administered by the United States Geological Survey. The site has therefore been the focus of numerous investigations including geophysical (Mewafy et al. 2011, 2013; Heenan et al. 2017), microbial (Bekins et al. 2001; Beaver et al. 2016, 2021), and geochemical studies (e.g., Cozzarelli et al. 2010). The site geology consists of ~20 m-thick moderately calcareous silty sand and outwash glacial deposits overlying clayey till of unknown thickness (Bennett et al. 1993). Figure 9.10c shows the MS measurements recorded at the site from 2011 to 2015. The initial MS results were reported in Atekwana et al. (2014). The data showed a marked increase (positive excursion) of MS within the zone of water table fluctuation which was not observed at uncontaminated locations. The MS decreases significantly within the saturated zone below the WTFZ. Microbial data from the site reported in Beaver et al. (2016, 2021) documented that this zone of positive MS excursions coincides with the methanic zone where active methanogenesis is occurring. The zone of iron reduction was above this enhanced MS layer representing a paradox. To explain this enigma, Beaver et al. (2021) suggested methanogens may switch their metabolism from methanogenesis to iron reduction resulting in the precipitation of magnetite.

Atekwana et al. (2014) suggested that MS measurements could be used as a low cost, rapid monitoring tool for assessing the extent of hydrocarbon contamination, and to delineate zones where magnetic mineral precipitation due to iron reduction was occurring. Nonetheless, continued measurements over a 4-year period (2011–2015) documented significant attenuation of the enhanced MS across the WTFZ by ~90%, suggesting that the MS signals were transient (Lund et al. 2017). Although the reduction in MS signal magnitude may result from dissolution of the magnetite (e.g., Ameen et al. 2014), recent studies by Ohenhen et al. (2022) provided evidence suggesting microbial-mediated anaerobic conversion of magnetite to maghemite.

The MS data across all three sites presented in the above case studies show MS excursions within the WTFZ consistent with observations reported at other hydrocarbon-contaminated sites (e.g., Rijal et al. 2010, 2012) and suggest that this interface is biogeochemically active, representing a hotspot of microbial activity. We also note that studies at a hydrocarbon-contaminated site in Iran by Ayoubi et al. (2020) further document the use of MS in predicting hydrocarbon levels of the impacted subsurface volume.

9.3.6 Temporal (Time-Lapse) Geophysical Investigations of Hydrocarbon-Contaminated Sites

One of the major advantages of the application of geophysical techniques to the characterization of LNAPL-contaminated sites is its ability to provide high resolution spatio-temporal images of the subsurface that can be used for monitoring natural attenuation as well as for the monitoring of active bioremediation. We provide some case studies below on the use of geophysics for monitoring LNAPL sites.

Sauck et al. (1998) provided one of the first documented case studies on the use of the self-potential method for mapping a LNAPL plume. In this study, self-potential measurements were first acquired in 1996 over the FT-02 plume at the WAFB site as described above in Sect. 9.3.1.1 (Fig. 9.11a). In 2007, the SP survey was repeated (Fig. 9.11b) and results were presented in Che-Alota et al. (2009). Both data sets were acquired at the same time of the year. The 1996 data (Fig. 9.11a) show a positive NW–SE trending positive SP anomaly reaching a maximum of ~24 mV coincident with the approximate plume boundary as delineated from hydrochemistry and GPR surveys (Sauck et al. 1998). This positive anomaly is in contrast with more negative SP values (around −12 to −30 mV) characterizing the background. Although the site was undergoing intrinsic bioremediation with large negative redox potentials, the source mechanism of the SP anomaly is attributed to diffusion potentials as the values are too small to result from a biogeobattery. In 2007, the repeated survey showed a similar positive anomaly characterizing the plume with two distinct differences: (1) the strength of the anomaly was attenuated and the maximum values are in the range of 3−6 mV; and (2) there is a region of negative SP anomalies (−18 to −34 mV) in the central part of the plume coincident with a soil vapor extraction (SVE) system that was installed in 2001 (Che-Alota et al. 2009). Thus, the decrease in the magnitude of the SP anomaly over the plume is attributed to ongoing natural attenuation processes as well as the reduction of the contaminant mass from the SVE system. Coincidentally, as detailed in Che-Alota et al. (2009), the geophysical anomalies (GPR, ER) observed in 1996 were significantly attenuated by 2007 with both the resistivity and GPR reverting to background conditions.

Fig. 9.11

a SP anomaly map of a portion of the FT-02 measured in 1996 and b in 2007. The grid showing the location of a soil vapor extraction system installed at the site in 2001 is superimposed on the 2007 SP map (modified from Che-Alota et al. 2009 with permission). The 2007 map shows significant attenuation in the magnitude of the positive SP anomaly compared to the 1996 map

Giampaolo et al. (2014) used the SP method to monitor a crude oil-contaminated site in Trecate, Italy. In 1994, an oil well blowout from the TR24 ENI-Agip operated exploration well resulted in 15,000 m³ of middleweight crude oil released to the environment contaminating soils and groundwater. Although cleanup efforts removed most of the contaminants, some of the crude oil infiltrated into the subsurface reaching the groundwater. The site geology is characterized by a thick deposit of Holocene glaciofluvial and fluvial deposits made up of poorly sorted silty sand and gravel in extensive lenses, typical of braided streams sediments (Giampaolo et al. 2014). The site has been the focus of several geophysical investigations (GPR, IP, ERT, SP) including Cassiani et al. (2014) and Godio et al. (2010). A time-lapse SP survey was conducted over a 12-month period (March 2010, October 2010, and March 2011), covering both contaminated and uncontaminated regions (Giampaolo et al. 2014). The corrected self-potential maps are shown in Fig. 9.12a–c and document significant time-lapse differences. In general, the SP anomalies are positive over the contaminated region for the March 2010 and 2011 data ranging from 10 to 65 mV, whereas the October data show a bipolar anomaly ranging from − 15 to 25 mV with the positive pole in the north and negative pole in the south. The authors relate the positive anomaly observed in the March 2010 and 2011 data to a decreased value of the electrokinetic coupling coefficient in the contaminated vadose zone because of biodegradation, whereas the bipolar anomaly in October is related to a difference in the redox conditions occurring in the northern and southern parts of the area. The small SP anomalies at this site rule out biogeobatteries as the source of the anomalies due to absence of electronic conductors.

Fig. 9.12

Self-potential maps obtained during March 2010 (a), October 2010 (b), and March 2011 (c) surveys from a crude oil-contaminated site in Trecate, Italy. Note the differences in the time-lapse images with the pronounced bipolar anomaly recorded in the October survey (modified from Giampaolo et al. 2014 with permission)

Incidentally, time-lapse SP monitoring of the Bemidji, MN plume described in Sect. 9.3.4 above by Heenan et al. (2015) using downhole SP probes imaged a strong dipolar anomaly with a negative anomaly above the WTFZ and a positive anomaly below. The observed SP dipole is centered around the zone of highest magnetic susceptibility within the WTFZ. Thus, the microbially-mediated magnetite layer serves as an electronic conductor, transporting electrons from reducing conditions (iron reduction and methanogenesis) below the saturated zone into the unsaturated zone where oxidizing conditions (iron oxidation) occur. This geobattery response appears to be transient in nature, probably driven by hydrobiogeochemical processes.

The above case studies illustrate that SP techniques can be used to not only map the extent of the LNAPL plume but can also map the dynamics of the plume driven by redox and hydrologic processes including monitoring the attenuation of the plume due to active and passive remediation. SP also has the advantage that it can be used to estimate the redox potential distribution (Eh) at organic-rich contaminated sites (e.g., Naudet et al. 2003, 2004; Abbas et al. 2017).

9.3.7 Other Emergent Geophysical Techniques

We would be amiss if we did not include other emergent geophysical techniques that can improve the delineation and quantification of LNAPL contamination.

9.3.7.1 Nuclear Magnetic Resonance (NMR)

Chapter 7 provides the principles of NMR technique and its application to determination of porosity and permeability. Here we provide an example of the use of NMR in the detection of NAPLs, as this technique has been extensively used in petroleum exploration to estimate hydrocarbon and water volume. In this field example, Fay et al. (2017) measured the spin–spin relaxation (T₂ relaxation) times and the diffusion coefficient (D) (D results not shown here) to investigate hydrocarbon impacted sediments at a site in Pine Ridge, South Dakota, USA. Oil field operations at the Pine Ridge, South Dakota site resulted in releases of petroleum products into the subsurface from multiple leaky underground storage tanks. Hydrocarbon contamination was first discovered at the site in 1992. Significant free product thickness of up to 4.3 m thick was observed in wells in 1995 when remediation efforts were initiated. At the time of the survey in 2016, 13–42 cm of free products was still present in the monitoring wells. The subsurface geology is characterized by lacustrine and fluvial sediments consisting of silts, silty sand with siltstone lenses. The water table elevation ranges from 7.8 to 11.1 m below ground surface with ~3.5 m of water table fluctuation. Supporting laboratory measurements were acquired to constrain the field data interpretation. Javelin and Dart NMR probes were used for data acquisition in two different bore holes.

Figure 9.13a shows the Javelin probe T₂ relaxation results. On this figure, warmer colors (yellow) indicate increased fluid volume at a particular decay time. The dashed white line on the plots is the T₂ at 0.1 s and is used to distinguish between the water and contaminant signal. Based on laboratory simulations, the contaminant in silt is expected to plot to the right of this line. The blue water symbol on the plots is the water table elevation at time of measurement and the grey water table symbols are the minimum and maximum water table elevations over six years. The results show bimodal T₂ distribution at several depths which is more pronounced at depths below the water table (below 8.4 m for MW-4 and 7.5 m for MW-16). Fay et al. (2017) interpreted the peak at long T₂ as resulting from the presence of contaminants. In both wells, T₂ > 0.1 s is observed at all measurements below the water table and within the smear zone. Although the authors did not consider the effects of biodegradation, we note that the water table fluctuation zone is the zone of most intense biogeochemical activity (often concomitant with mineral precipitation such as magnetite as well as biofilm formation) with maximal changes in geophysical signatures (e.g., Werkema et al. 2003; Atekwana et al. 2014; Sharma et al. 2021). Figure 9.13b shows a plot of the fluid content estimated from the T₂ logs in MW-4 and MW-16. The blue line is the total fluid content, and the red line is attributed to contaminant. In MW-4, the total fluid content ranges from 12 vol% in the unsaturated zone to a maximum of 28 vol% within the saturated zone. Similarly, the contaminant volume is lower in the unsaturated zone (1 vol%) and increases to a maximum of 5 vol% within the saturated zone. In MW-16, the contaminant volume increases to 9.5 vol%.

Fig. 9.13

a Nuclear magnetic resonance measurements of the T₂ logs measured with the Javelin in 15 cm (6 in) wells MW-4 and MW-16. The color shows the distribution of T₂ values, with warmer colors (yellow) indicating higher signal amplitudes. The depth of the water table in March 2016 is indicated in blue; the gray symbols show the range of water levels observed between 2010 and 2016. The dashed white line shows the T₂ = 0.1 s line. b Total fluid content (blue) and fluid content with long T₂ (red). MW-4 and MW-16 are the monitoring wells. Solid black line is the water table elevation at time of measurement and dashed black line shows the range of water levels measured from 2010 to 2016. Figure modified from Fay et al. (2017) with permission

The observations documented in the Fay et al. (2017) study suggest that NMR is an effective tool that can be used to detect the presence of contaminants and quantify their volume in the sediments. Although not discussed here, the NMR tool can also be used to detect in situ biofilm formation expected to be associated with biodegradation especially within the water table fluctuation zone (e.g., Kirkland et al. 2015).

9.4 Conclusions and Key Take-Aways

We have provided case studies from environments across the globe that show that geophysical investigations can be used to detect fresh spills, as well as effectively used to assess the biogeochemical changes occurring from intrinsic and engineered bioremediation at contaminated hydrocarbon spill sites. In fresh or new spills, the geophysical response is driven by the displacement of (and mixing with) conductive pore water by the resistive LNAPL resulting in a resistive response imaged by electrical geophysical techniques. However, with time, biodegradation causes more conductive pore waters from ionic metabolic by-products resulting in a conductive response which is by far the most common response observed at hydrocarbon sites. This makes electrical geophysical techniques (ERI and GPR) the most applied techniques. Nonetheless, not all biodegraded sites show a conductive response as the response is also dictated by the background geology and fluids properties. Where clays and saline waters characterize the background conditions, then even conductive pore waters resulting from biodegradation would result in a resistive response. Thus, knowledge of the background conditions is important for the investigation and interpretation of the geophysical signatures.

The geophysical response is also driven by microbial-mediated redox processes that result in the precipitation of minerals. The choice of geophysical technique to use to delineate these “hot zones” of microbial activity depends on the contrast in physical properties between the biomineral and the mineralogy of the background geology. We provided case studies showing that magnetite which can be precipitated during microbial-mediated iron reduction is both magnetic and conductive and can be imaged using magnetic susceptibility or SIP techniques. Interestingly, the magnetic susceptibility response was strongest within the water table fluctuation zone pointing to this zone as the most biogeochemically active. In fact, it is within this zone that studies have documented the strongest geophysical responses.

The geophysical detection and monitoring of hydrocarbon contamination are not trivial, as results are often mixed and data interpretation must proceed with caution. For example, the application of the same geophysical technique at different locations at the same site can produce dramatically different results due to the variability of the contaminant mass distribution and the activity of indigenous microorganisms. As depicted by the magnetic susceptibility response in Fig. 9.10, the largest geophysical response may occur within the water table fluctuation zone, where free phase LNAPLs were found, resulting in an abundance of electron donors. This is not surprising as enhanced biogeochemical cycling driven by dynamic environmental conditions results from the mixing of electron donors and acceptors stimulating microbial respiration and shifts in microbial activity, carbon turnover, and elemental cycling. However, the magnitude of the geophysical response decreased downgradient away from the core of the plume where free phase LNAPLs existed to distal ends of the plume where the plume was in the dissolved phase, as well as within the saturated zone.

As previously suggested by Atekwana and Atekwana (2010), the geophysical response of hydrocarbon-contaminated media depends on several factors including: (1) the type of the contaminant present (crude oil, jet fuel, diesel fuel, etc.), (2) their distribution and partitioning into different phases (vapor phase, residual and entrapped phases, free phase, dissolved phase) in the unsaturated and saturated zone, (3) dynamic hydrologic processes (e.g., advective transport, seasonal recharge), (4) release history (e.g., continuous release over a long time versus a single release), (5) the saturation and wetting (oil or water wetted) history of the contaminated media, and (6) biological processes. We add to this list (1) background geology (e.g., clay vs sand), (2) pore water salinity, (3) hydrobiogeochemical processes, and (4) time of year when measurements are made. In conclusion, geophysical technologies offer clear benefits in the characterization of hydrocarbon-contaminated sites especially when optimized by biogeochemistry data. They can therefore be used for identifying contaminated areas, remediation monitoring, and post-remediation monitoring.

Looking to the future, more studies are needed in the following areas:

• 1. Improved understanding of how hydrobiogeochemical processes drive the geophysical response. For example, the transient magnetic susceptibility is curiously related to the hydrology and water levels at the site and appears to be at a maximum during periods of high-water level but declines during drought periods. Thus, a mechanistic understanding is needed to explain the transient magnetic susceptibility signatures.

• 2. Relating the geophysical signatures to terminal electron acceptor processes. At present, only responses related to iron reduction can be conclusively determined.

• 3. Estimating the volumetric content of the LNAPL from the geophysical signal. There is some encouraging evidence provided by Deng et al. (2018) that SIP can be used to estimate the volumetric content of the LNAPL. However, this study was a controlled spill in a well-characterized sandbox experiment. To this last point, we need more field investigations that address this point as well as determine the threshold volume that can be detected by geophysical methodologies, especially in complex geology. NMR may hold such potential as it can both detect and quantify the LNAPL volume.

• 4. We need more innovations to make SIP instruments more readily usable in field settings like current ERI technologies.