The ability to characterize the geometry and lithology of Quaternary sediments is important to scientists who investigate groundwater movement, geoarchaeology, materials prospecting (e.g., gravel), environmental contamination and remediation, and paleoenvironmental studies. Often these studies are restricted by the limited information attainable via traditional geomorphological techniques. While there are geophysical methods for gaining information about the near-subsurface, such as ground penetrating radar (GPR) or shallow seismic surveys, they only function well under select conditions. Electrical resistivity imaging (ERI) can quickly produce high-resolution images of the shallow subsurface under many field conditions. ERI measurements work well in both resistive sediments, such as gravels and sands, as well as conductive sediments like silt and clay. Resistivity is an inherent property of all materials, and it measures the degree to which a material resists the flow of electrical current. If a current is introduced into the ground, the resulting electrical field can be measured. Thus, a two-dimensional cross section can be produced showing the resistive properties of a sediment package several meters behind an exposure. This aids in the interpretation of the material and structural features that may be present but not exposed. This methodology is successful in imaging some subsurface architecture, but there are limitations to the resolution of the surveys. ERI, when integrated with detailed geomorphologic analysis, provides enhanced insight for inferring the processes of sediment emplacement and deformational processes.
Continued efforts to map the geometry and lithology of Quaternary sediments by geomorphologists and surficial geologists have been hampered by an inability to easily obtain information about the shallow subsurface. Mapping these sediments is important because they contain economically significant sources of construction aggregate and groundwater as well as information about past environments. Shallow subsurface geophysics has provided new ways to obtain information about subsurface architecture in Quaternary sediments.
Data from shallow seismic or ground penetrating radar (GPR) and more traditional methods like coring have been used to produce models of the subsurface and act as a guide for subsequent coring surveys. However, each of these methods has limitations when used in mapping Quaternary sediments. While cores can accurately represent packages of sediment in a single dimension, detection of lateral changes is not possible without extensive coring programs. As Quaternary sediments are often complex, laterally changing over short distances, there is a need for more detailed information across an entire site. Shallow seismic equipment is bulky, heavy, and expensive for this type of fieldwork and cannot image near-surface (upper 2–3 m) materials. GPR has attenuation problems in sediments composed of clays and silts. The electromagnetic waves sent into the ground attenuate more in these conductive materials, and tend to only send weak signals back to the receiver (Moorman, 1990; Reynolds, 1997). Electrical resistivity imaging (ERI) works effectively in sand and gravel, as well as in fine sediments like silt and clay, since it measures the resistance of the material to electrical conduction and not the reflectance of electromagnetic waves.
The purpose of this paper is not to infer the paleoenvironments, but to show the utility and application of ERI in acquiring more information that can be used in interpreting site formation processes. Since geological structures by nature are 3-dimensional (3-D), incorporating data in a third dimension from a geophysical survey should provide useful information about the sedimentary architecture and aid in the subsequent interpretation of a glacial history for an area.
Three sites from an area of hummocky terrain located around McGregor Lake, Alberta, Canada, were selected for this study (Fig. 1). Hummocky terrain is composed of hills and depressions of variable size and shape that occur in a previously glaciated region (Munro and Shaw, 1997). This type of terrain has been recognized extensively in Europe and North America, and it has been attributed to many types of formation processes (Gravenor, 1955; Stalker, 1973; Stalker, 1977; Benn, 1992; Munro-Stasiuk, 1999a; Boyce and Eyles, 2000; Bennett, 2001; Evans et al., 2006). The sites were chosen because hummocks in these areas have, arguably, the best exposures worldwide (Munro-Stasiuk, 1999b), which allow the stratigraphy to be matched to the ERI model generated in the survey. These sites all have large packages of fine-grained sediment, often comprising a majority of a site, which makes GPR ineffective. Extensive fieldwork has also been conducted in this region by others (Munro and Shaw, 1997; Eyles et al., 1999; Munro-Stasiuk, 1999b, 2003; Smith, 2006). The reservoir provides extensive shoreline erosion, creating excellent outcrop exposures.
BACKGROUND AND METHODS
Electrical resistivity is a measure of how difficult it is to pass an electric current through a volume of material with a given length and cross-sectional area (Ward, 1990; Reynolds, 1997; Loke, 2001). Electrical resistivity of sediments or rock is measured in Ωm (ohm • m) and is a function of porosity, saturation, resistivity of the pore fluids and the solid phase, and the material texture. Previous uses for ERI have included: archaeological surveys (Noel and Xu, 1991); mapping of bedrock channels by Chen et al. (1996), Ramage et al. (1998), and Gilson et al. (2000); mapping of subsurface pipes (Vickery and Hobbs, 2002); remediation site analysis and mapping (Bentley and Gharibi, 2004); and ground ice prospecting (Lugon et al., 2004).
The resistivity (ρ) of a volume of material can be calculated by the equation: ρ = R(A/L), where R is the electrical resistance, A is the cross-sectional area, and L is the length. The measurement is reported as apparent resistivity (ρa): ρa = k·(▵V/I), where ▵V is the voltage difference (measured in volts), I is the injected current, and k is a geometric factor based on the electrode arrangement. The geometric factor incorporates the distance from each current electrode to each potential electrode and a half-space. Surface ERI surveys are conducted at the boundary of two semi-infinite media, the earth and air (Fig. 2). The resistivity of air is considered infinite, so it does not contribute in the electrical current flow. The earth materials, however, are conductive, and therefore the lower semi-infinite medium, or the half-space earth, conducts all the injected current (Gharibi, 2005, personal commun.). The half-space is included to help model the flow of electricity more accurately downward and outward from each of the four electrodes in the system. This only holds true if the half-space is homogeneous. Since the half-space is usually heterogeneous, the measurements taken in the field are considered to be apparent resistivity (ρa) (Baines, 2000).
ERI operates by applying a current through two current electrodes and measuring the resulting voltage difference between two potential electrodes. Electrical resistivity imaging combines collected apparent resistivity measurements to create pseudosection profiles (contoured resistivity value diagrams) of the subsurface. The arrangement of the four electrodes, known as an array, affects the depth of investigation, sensitivity, resolution, and the incorporation of noise into each apparent resistivity measurement (Fig. 3). The Wenner array is robust in the presence of measurement noise and is well suited to resolving horizontal structures because it is more sensitive to vertical changes in resistivity than to horizontal changes (Ward, 1990; Loke, 2001). The dipole-dipole arrangement is very sensitive to horizontal changes in resistivity. This makes the dipole-dipole array well suited for imaging vertical structures. Variations in the apparent resistivity represent different sedimentological conditions (Baines, 2000; Baines et al., 2002). Because tills, fluvial and lacustrine sediments, and bedrock are expected to exhibit large contrasts in such properties, electrical resistivity should be well suited to resolving the subsurface architecture of unconsolidated sediments in hummocky terrain.
We conducted ERI surveys using a 56 electrode AGI Sting/Swift R1 IP earth resistivity meter (Fig. 4). Once the survey had been completed, the survey lines were processed using RES2DINV (Loke, 2002) to create pseudo-section profiles. The numerous measurements taken in the field along a line yielded a contoured distribution of apparent resistivities for the subsurface. The apparent resistivities were then converted to true resistivity values using an iterative numerical inversion that determines a least squares fit to the measured data (Ward, 1990; Loke, 2001). The inversion searches for the smoothest possible resistivity values that would produce the same apparent resistivity gathered in the survey.
Because material texture plays an important role in the resistivity of a package of sediment, particle size analysis using laser diffraction was completed on the sediments from each site. While there are other contributing factors to the resistivity of sediment at the sites, the surveys were done in a short period of time so as to minimize the chance of any change in moisture content of the ground. There was also no significant chemical input, such as fertilizers, that may have affected the results at any of the sites.
The field sites are discussed separately and include a description of each site, the sediment composing the site, the ERI survey conducted, and the interpretation of the ERI profiles.
This west-facing exposure (250°) (Fig. 5) is 140 m in length and is composed of two small hummocks (3–4 m in height) and one large hummock (∼10.5 m in height). The area around this site is mainly hummocky terrain, with some flat areas between hummocks. The surface of this site is sparsely covered in cobbles and small boulders. The exposure is dominated by diamicts. The first diamict (1), which makes up most of the site, is stone-poor (∼2%) with a silty–fine sand matrix. The matrix is massive, although some sections that are layered have weak, undulating bottom contacts (Smith, 2006). These layers of diamict dip slightly (∼5°) to the north. Some soft sediment clasts or massive silty inclusions are incorporated into the diamict, and large sections are subvertically fractured. A second, stone-poor (∼1%) diamict (2) with a more sandy matrix intrudes (apparent dip of 33°) into the primary diamict. Both diamicts are capped with a variable-thickness (but no larger than 90 cm thick) layer of silty–fine sand diamict, which has a higher clast content (and larger clasts, some up to 10 cm) and darker color than the first diamict. It has sharp contacts with the surrounding diamict and is massive in appearance. Horizontal and dipping beds (up to 40°) of medium to coarse sand (3) comprise the majority of the largest hummock. This section has few clasts and none larger than 3 cm. Dipping beds contain small faults and slight offsets. Although it is suspected that a high-angle reverse fault, or several faults, cut through the blocks of angled beds and horizontal beds, they are only inferred based on the larger angled blocks and are not visible. The sands are topped with a 1-m-thick layer of well-sorted, bedded gravels (4) that dip at 40°. Beds are typically 5–6 cm thick, although some beds are larger. Many of the beds show offsets of 1–2 cm. The gravels are cut by a high-angle reverse fault (74°) along the north side, and they pinch out along the south side of the hummock.
This site was surveyed using two ERI lines parallel to the outcrop. Both lines were 98 m in length and used a Wenner array with a 1 m electrode spacing. The lines were 14 and 18 m from the outcrop edge to ensure that edge effects were minimized.
The first line (Fig. 6A) has detected the contact between the gravel-sand and the diamict, but the high-resistivity zone is completely missing in second survey line (Fig. 6B). It seems unlikely that this is the result of edge effects or noise, based on the distance away from the outcrop edge. It is then likely that the contact itself strikes at a rather dramatic angle—at least a 15° change in direction from the first survey line to the second (since they are only 4 m apart). This shows that the sand and gravel unit is not consistent throughout the hill. The contact of the sand and gravels with the surrounding till is sharp and compact and almost vertical. The small zone of slightly more resistive material (35–40 Ω·m) under electrodes 56–60 (Fig. 6A) and the zone of slightly more resistive material (35–60 Ω ·m) under electrodes 60–65 (Fig. 6B) could be interpreted as the same sandier diamict striking at ∼115°. The sandier diamict may also not be a consistent thickness throughout the site. Only the larger architectural features were imaged, and subtle textural changes may be below the detection level.
This site (Fig. 7) is 60 m in length and faces east (95°). This section has some moderate topographic relief, as the site dips in the center into a small depression. The surface of the hummock slopes away from the outcrop toward the west into a small gully. The site is sparsely covered in boulders. Concentrations of boulders appear where two gravel layers are close to the surface at either end of the outcrop. The majority of the hummock is composed of planar beds of clay, silt, and fine to medium sands (1), which dip (∼45°) to the south. Clasts are infrequent, and none are larger than 5 cm. Small flame structures are visible in some layers. Bedding becomes more convoluted and has a higher dip angle closer to the thrust fault (∼45°). North of the fault is a silty–fine sand diamict (2) (similar to the diamict that composes the majority of site 1). Clasts (<3%) are up to 15 cm in size, but are dominantly 1 cm. There is very weak bedding visible in some areas of this unit oriented downward slightly to the north. Some lower contacts are uneven and slightly undulating, while other contacts are sharp. The site also contains large massive gravel lenses (3). They are moderately to well sorted, subrounded, and range from 1 to 20 cm in diameter but predominantly are 1–2 cm in size. The matrix is fine sand with some silt. The gravels are a mix of both matrix-supported and clast-supported cobbles and pebbles and are moderately sorted, but tend to become more poorly sorted toward the top of each lens. Contact with the surrounding diamict above the lenses is uneven in some places and erosional along the bottom contacts.
The hummock was surveyed using three 56 m Wenner array lines with 1 m electrode spacing. As with site 1, the first line was placed 10 m in from the outcrop edge. The following two lines were placed 14 and 18 m from the outcrop edge.
In all three ERI profiles (Fig. 8), the gravel beds can be seen (high resistivity—reds), and they appear to be consistent in both slice A and B and 10 m and 14 m from the outcrop. Sand, as a more resistive material, is expected to have a higher resistivity, but the image clearly shows a more conductive body (blues). There are two potential reasons for this. First, there is a gradational change going into the site from the sandy layers found at the outcrop to something more conductive, such as clays and silts. This is contrasted with the more resistive diamict (yellows and greens), which contains more silts and clays. Second, the silty–fine sand diamict that overlies the bedded sands and silts strikes into the site toward the southwest (∼220°). Facies D continues to overlie the bedded sands and silts; it just lies overtop them further along the survey line. The conductive body is inconsistent through the site as well. It shrinks slightly in slice B, and decreases again in slice C, where it is separated into smaller blocks with more resistive material between. The second explanation gives the potential for deformation, or brecciation and incorporation of the bedded sands and silts seen in slice C. Slice C also shows the imaging of more resistive units near the surface at electrodes 24 and 33. While the surface was rocky, and driving the spikes into the ground was sometimes difficult, it is unlikely that these more resistive sections would be large boulders, instead they are most likely units of more resistive sands and gravel.
This site (Fig. 9) is 225 m in length, and the exposure faces east toward site 1 on the east shore of McGregor Lake. This site has the most dramatic topographic relief of any of the study sites. It is a large hummock, and the apex is 20 m above lake level. The surface of this section is sparsely covered in boulders. It should be noted that recent slumping has masked large portions of the hummock, making this site ideal for testing ERI as a reconnaissance tool. The site is composed of several stone-rich (∼10%–12%) silty–fine sand diamicts, interbedded clays, silts, and fine sand layers, and a stone-poor (∼2%) silty–fine sand diamict. Two silty–fine sand, stone-rich (up to ∼20%) diamicts partly cap the site (1). Each diamict is ∼1 m in thickness, and the top of the diamict is slightly lighter in color. Most of the clasts are less than 5 cm, but a few are up to 75 cm. There are several spots within both diamicts where small lines of boulders dip at ∼25° toward the north. The lower, darker-colored diamict also contains sporadic, small gravel lenses or inclusions often clustered near some of the more massive boulders. A large section of silty–fine sand diamict containing soft sediment clasts (2) (often up to 1 m in size) has been deposited on the north side of the hummock. The soft sediment clasts are often a clayey–silt diamict, but they can also be bedded, fine sand, or thin gravel lenses. The soft sediment clasts are often inclined, and, in the case of the sand beds, have bedding preserved. The silty–fine sand diamict is stone-poor (∼1%). The center of the hummock consists of inclined beds dipping 35° toward the north end of the outcrop. Individual beds consist of massive, silty diamict (3), ∼3 m thick, and overlie another diamict with richer clast content (2%–3%) about a meter in thickness below. Below this is a silty diamict with almost no clasts, ∼1.5 m in thickness. This is overlying a layer of interbedded silt, fine and medium sand similar to those beds of interbedded sands and silts in site 2, only not disturbed; the beds have just been reoriented. A thrust fault at the base of this unit dips at 35°, the same angle as the beds above. Contacts between the bedded units are sharp at some bed contacts and erosional at others. Below the fault are bedded layers of silt and fine sand (4) with very few clasts (<1%), again similar to those found in site 2. These layers have been deformed and curve upward toward the fault. The south end of the hummock contains a silty–fine sand diamict with few stones (5) (<1%).
The hummock was surveyed using two 112 m Wenner array lines with 2 m electrode spacing. As the system is restricted in the number of electrodes, a 2 m electrode spacing was selected so as to cover a larger portion of the site and image deeper into the subsurface. Each electrode was then shifted 1 m along the line, and the survey was repeated. Both sets were combined into one profile to give the same resolution as a survey with a 1 m electrode spacing, but giving a greater depth of investigation, since the line has a 2 m electrode spacing. The first 112 m line was placed 14 m from the outcrop edge, and the second 112 m line was placed 18 m from the outcrop edge. The survey for this site also included two 56 m dipole-dipole array lines with a 1 m electrode spacing. These shorter lines were placed at 10 m from the outcrop edge and 22 m from the outcrop edge and centered at the midpoint of the two longer lines.
The stony diamict was imaged along the north side of the line (Fig. 10), between electrodes 77 and 112 (high resistivity values—oranges and reds). It has an average thickness of ∼2 m. Another area along the south side of the line, from electrodes 1–29, also contains high-resistivity values and has an average thickness of 1.5 m. The area of low resistivity, from electrodes 13–29 and 71–97 underneath the stony diamict, is interpreted to be the same or similar stone-poor, silty–fine sand diamict expressed all over the area. The U-shaped zone (electrodes 31–69) of medium-value resistivities (green and yellow) is interpreted as the bedded layers of fine sand and silt that have been faulted. All slices from this site are displayed in Figure 11.
DISCUSSION AND CONCLUSIONS
There are both limitations and advantages to using ERI as a research tool. ERI is a relatively cheap and efficient way of gathering subsurface images and the equipment is less costly and bulky. The survey itself is noninvasive, meaning there is no noise, explosives, ground disturbance, or waste. This mitigates some of the problems when working on private land. ERI can work in areas where seismic or GPR may not be useable or cannot gain usable information, such as those sediments with fine particle sizes. Estimated strikes and dips of contacts can be acquired from these contrasting materials in some situations. At site 1, an estimate of the strike of the sand-gravel contact with the adjacent silty–fine sand diamict can be gained from the sharp resistive boundary (the compact contours). A second estimated strike can be gained from the same survey for the sandier, intruding diamict. While the contacts with the sandier diamict were not imaged as clearly as other contacts, the center of the sandier diamict's ERI signature on both profiles gives a rough estimate of the strike of the body. The interpreted geometry of the sediment packages can give more information about the processes of deposition or deformation that occurred at the site, leading to better interpretations of site history. ERI also has the ability to establish the contiguity of contrasting packages of material, which can be extremely important for the interpretation of a site's history. Intraclasts can play a large part in the consistency of resistivity for Quaternary sediments. A particular type of diamict may have regions with more clasts than other areas where that diamict is expressed. This would show up as a slightly higher resistivity than the overall values for that diamict. At site 2, there is a lack of consistency from the outcrop directly inward in relation to the bedded silts and sands. The ERI signatures of these sediments at this site shift position and become incorporated within the silty–fine sand diamict, which strikes into the hummock. Site 3, on the other hand, displays consistency of features through multiple survey profiles. This can be seen in the expression of the stony diamicts at the surface in both the long and short profiles. This is also expressed in the consistent shape of the hummock architecture presented in the profiles.
ERI also makes a useful reconnaissance tool. While this study did not incorporate coring, ERI gives more information about the subsurface away from an outcrop. This gives the researcher a better idea of where they may wish to do a sample core, which in turn would enhance the ERI information further if that core were taken along the line of an ERI profile.
The most obvious drawback to ERI investigations is the inability to detect fine-scale details like thin interbedding material such as those interbedded clays, silts, and sands at site 2 or the contacts of layers of the same material, such as the layered silty–fine sand diamict at site 1. These types of details become lost in the bulk measurement and averaging of resistivity over an area. While intraclasts can affect the resistivity of a package of sediment, the intraclasts themselves are quite often not imaged. They are, instead, averaged into that bulk resistivity. They could potentially offset the resistivity of a package of sediment to mask it as another package of sediment at that site, if the resistivity values of those clasts were contrasting enough and in sufficient quantity. An example of this would be a diamict with a conductive matrix but with a higher stone content. The stones increase the bulk resistivity, giving it a higher resistivity value. This value may coincide with another package of sediment that has a different material texture altogether. ERI pseudosection profiles also have an inability to image sharp sedimentary contacts. Since the profile is contoured, resistivity boundaries must be interpreted based on standard ranges of sediment resistivity. As there are different properties that can affect resistivity from site to site, such as pore water content and geochemical characteristics, similar sediments at two different sites may have differing resistivity values when compared. For example, if one site has higher pore water content or anthropogenic chemical input (e.g., fertilizers) than the other site, resistivity signatures will differ significantly, making intersite comparisons difficult. This may be as simple as it having rained heavily the previous night at one site but not at the other before a survey was done. This makes intersite comparisons difficult if survey areas are far apart or if surveys are conducted at different times of the year. Problems can also arise with survey design. If a survey line is too close to an outcrop edge, strong resistive bodies may develop where there is no resistive feature in the subsurface. This edge effect can create misinterpretations about the subsurface. ERI cannot be used as a stand-alone survey tool. It must be used in conjunction with traditional geomorphological techniques. But with traditional field techniques, ERI can be an indispensable tool for gaining more information about a site by giving added information in a third dimension.
This project was funded through the support of Natural Sciences and Engineering Research Council of Canada (Discovery Grant) and the University of Calgary, University Research Grant held by D. Sjogren. We would like to thank our field assistants and Brenda Mottle for all the technical support.