The Use of Geophysical Methods to Characterize Hydrogeologic Systems Across Multiple Scales
Geophysical methods are seeing increasing use as a means of characterizing hydrogeologic systems. Captured in the geophysical data is information about the large-scale structure of the subsurface, and information about the smaller-scale properties that govern groundwater flow and contaminant fate and transport. At all scales, various geophysical data sets - field data and laboratory data - inform our understanding of the link between geophysical data and the hydrogeologic structure and/or properties of interest. At the field-scale, collaborations between hydrologists and geophysicists allow us to explore the link between the structure seen in the geophysical images and the structure relevant for modeling flow and transport. A critical part of such work is developing ways to quantify the uncertainty in the structural models derived from the geophysical data. At a smaller scale, we can extract from geophysical images information about the spatial heterogeneity of a field site. But we need, again, to better understand the link between what is seen in the geophysical images and what is relevant for hydrogeologic applications. At the laboratory scale, and below, there is evidence that geophysical properties are sensitive to a wide range of chemical, biological, and physical properties. New observations are being made in laboratory studies that suggest exciting new potential applications of geophysical methods. The challenge is to understand how the relationships observed at this small scale, between the geophysical properties and the chemical, biological, physical properties of a material, can be upscaled for the interpretation of field-scale geophysical data.
Chargeability and the Polarizable Volume Fraction.
A new relation between chargeability and the volume fraction of material polarizable by electrode polarization is
derived and supported with both published and new experimental evidence.
Approximate empirical relations have been presented before, for example, between the volume fraction, φ,
and the Newmont 331 chargeability, m331 = 2φ, and between the Percent Frequency Effect (PFE) and
the volume fraction PFE=3φ. Here, I present a simple theory that uses the Hashin-Shtrikman equations for
the bulk resistivity of a two phase system consisting of a uniform host and resistive (polarized) grains at low
frequency and conductive (unpolarized) grains at high frequency. The chargeability in V/V and the volume
fraction are found to be related by,
m=9φ / 2+5φ+2φ2.
Or, for small volume fraction,
The chargeability is shown to be independent of grain shape, size and composition, as well as pore water chemistry and concentration over a wide range of these variables, so that the volume fraction can be recovered from the measured chargeability under a wide range of conditions. The resistivity of the host can be recovered from either the low frequency limit or the high frequency limit of the bulk resistivity, once the polarizable volume fraction is known.
Heavy Mineral Provenance of Late Jurassic to Early Cretaceous Reservoir Sandstones in the Flemish Pass and Orphan Basins, Offshore Newfoundland
The Orphan and Flemish Pass basins are frontier offshore deep water basins, located approximately 400 km northwest of the Avalon Peninsula, Newfoundland. Detrital heavy minerals from potential oil-and gas- reservoir intervals (Late Jurassic to Early Cretaceous sandstones) in three industry exploratory wells (Mizzen L-11, Baccalieu I-78 and Blue H-28) were studied for provenance analysis. Three heavy mineral approaches were used to determine provenance and make correlations: (1) U-Pb geochronology of detrital zircons, (2) detrital heavy mineral grain counts and ratios, and (3) geochemistry of detrital tourmalines. U-Pb Concordia ages of 40-85 zircons per sample were determined using LA-ICPMS at the INCO Innovation Centre Laboratory at The Memorial University of Newfoundland. Mineral identification, grain counts, and qualitative grain analyses have been made in a semi-automated fashion using a scanning electron microscope (SEM) equipped with and energy dispersive x-ray (EDX) spectrometer mineral liberation analysis (MLA) software. Detrital U-Pb geochronology, in conjunction with other lines of evidence, shows that a change in sediment source occurred in the Northern Flemish Pass basin during the Late Jurassic to Early Cretaceous North Atlantic rifting stage. During the late Jurassic, a mixture of first-cycle Mesozoic, Paleozoic, Early Neoproterozoic, and second-cycle Mesoproterozoic and Paleoproterozoic sediment sources dominated; whereas in the Early Cretaceous, second-cycle Proterozoic sources were dominant. This change in sediment sources is interpreted to represent a switch from sediments derived from distal hinterlands with significant Paleozoic crystalline basement (such as the Avalon Uplift, Central Mobile Belt, or Variscan Iberian Massif) to locally derived sediment (Avalon zone basement). In the West Orphan basin, a mixture of first- and second-order Paleozoic, Early Neoproterozoic, Mesoproterozoic and Paleoproterozoic sources appear to dominate in the Early Cretaceous (Albian). Sediments in the Cretaceous interval are interpreted as being derived from a north-western or western hinterland (consistent with a source of mixed Central Mobile Belt, Avalon and Grenville affinities).
A Differential Interpretation of the Cementation Exponent
Between 1950 and 2002 the total volume of reserves discovered has run to over 1500 Bbbl. for oil and 7.5 Tcf.
for gas. Over half of these resources has already been produced, and has driven the global economy for the
last fifty years. All of the assessments of the volume of hydrocarbon reserves were made using Archie's
relationships (1942). It would be difficult, therefore, to overestimate the impact of either the petrophysical
techniques or Archie's relationships on the worldwide economy. Archie's laws link the electrical resistivity of a
rock to its porosity, to the resistivity of the water that saturates its pores, and to the fractional saturation of the
pore space with the water, and are used to calculate the hydrocarbon saturation of the reservoir rock from
which the reserves are then calculated. Archie's laws contain two exponents, m and n, which Archie
called the cementation exponent and the saturation exponent, respectively. The conductivity of the hydrocarbon
saturated rock is highly sensitive to changes in either exponent. However, despite the importance of the
cementation exponent, few petrophysicists, commercial or academic, are able to describe its real physical
meaning. The purpose of this contribution is to investigate the elusive physical meaning of the cementation
exponent. We review the traditional interpretation of the cementation exponent and consider the extension of
Archie's first law to two conducting phases. Consequently, we develop a new differential interpretation of the
cementation exponent that is based on a new definition for the connectedness of the conducting phases in a
porous medium. In this interpretation the connectedness of a porous medium is defined as the availability of
pathways for transport, where the connectedness is the inverse of the formation resistivity factor, G =
σo/σw = 1/F (and may also be called the conductivity formation factor). Porosity is
defined as the fractional amount of pore space in the usual manner. Connectivity is defined as the measure of
how the pore space is arranged, is given by χ = φm-1, and depends upon the porosity and
the cementation exponent m. The connectedness is then given by G = χ φ, and depends upon
the amount of pore space (porosity) and the arrangement of the pore space (connectivity). The rate of change
of connectedness with porosity dG/dφ = mχ depends upon the connectivity χ and the
cementation exponent m. Hence, the cementation exponent can be interpreted differentially as the rate of
change of the connectedness with porosity and connectivity, m = d2G/dχdφ.
Multi-dimensional, Multi-variable, and Multi-scale Modeling of Reservoir Heterogeneities
Seismic wave velocities and energy attenuation are important seismic attributes used to assess reservoir physical properties. However, reservoir properties are often characterized by dissimilar 3D spatial distribution at various scales. For example, the heterogeneity of elastic properties at the scale of 10 ∼ 100 m influence the seismic image of geological targets whereas the heterogeneity of permeability at the scale of <10 m affects the recovery of oil. In order to obtain accurate seismic attributes, a model is required to account for the impact of both large- and small-scale heterogeneity on the transmission of seismic energy. A new approach is developed to model 3D reservoirs that capture different scales of heterogeneities and simultaneously honor statistical features observed in the well logs. Our approach provides a platform to estimate rock physics properties at the scale of well logs and to study the behavior of seismic waves in reservoirs with various scales of heterogeneities. This technique is applicable to various types of heterogeneous reservoirs and can also construct models to study other physical properties such as effect of heterogeneous distribution of conductivity on propagation of electromagnetic waves (not included in this study). The algorithm is implemented under both WindowsR and Linux environment using C/C++ and is adapted to distributed-memory, multi-computer cluster. The performance of the sequential and parallel codes is analyzed. As a case study, we applied our model, combined with the Biot-Gassmann theory, well logs from one borehole, and acoustic impedance inversion from surface seismic investigation, to the Mallik permafrost gas hydrate research site in the Canadian Northwest Territories. We find that the degree of lateral heterogeneity play a significant role in assessing the amount of gas hydrates and thus more adequate lateral sampling is desired to constrain the resource assessment in the reservoir. The results suggest the total hydrate amount is nearly an order of magnitude lower than early estimates in which small-scale heterogeneities were not accounted for.
Determination of the moisture and grain-size distribution within a waste rock pile using induced polarization
Nickel concentrations above the allowed norm have been measured in water samples in the vicinity of the Petit-Pas waste rock pile at the Tio mine. In order to understand the chemical and hydrodynamic mechanisms responsible for the Ni dissolution and release in the environment a large hydrogeological/geotechnical/geophysical study was undertaken on the very large rock pile (600 m x 300 m x 150 m). Hydrogeological and physical properties were measured in the laboratory on samples of waste with various moisture content and salinity. Induced polarization allowed the determination of electrical resistivities and chargeabilities that were to be used to interpret time-domain IP tomography surveys carried out on the waste rock pile. Simple petrophysical models such as Archie or Schon models, with a resistive matrix and a conductive electrolyte, are not appropriate since the waste material consists of up to 70 percent of hemo- ilmenite, a conductive and magnetic ore. IP survey data were inverted in 2-D and 3-D. The resistivity models for all survey lines show common features. From surface to 25-30 m depth, the subsurface is resistive and show lateral changes which suggests large grain size heterogeneities and low water content; below and up to 50 m, a quasi uniform layer with resistivities in the range 30-100 ohm.m suggests fine-grained material with increased saturation. Finally below a depth of 50 m, the subsurface becomes very resistive again, indicative of coarse low-moisture content material. The chargeability model shows no structural correlation with the resistivity model. In general, chargeabilities are very large (up to 200 ms) and show no layering. From the laboratory and the survey results, we interpret the resistivity to be sensitive to water content and salinity while chargeability is sensitive to the metal concentration.
Gravity and Magnetic forward modeling for Complex 3D Geological Bodies
In geological and geomechanical modeling, tetrahedral grids are very successful to describe subsurface features In particular, tetrahedral grids provide unprecedented accuracy for modeling heterogeneous complex geological structures; Furthermore, the tetrahedral mesh resolution can vary in space, to conform to subsurface heterogeneity. The resulting models then contain fewer elements than prismatic of Cartesian grids which have a fixed resolution. We use tetrahedral grids to describe 3D geological structure and to calculate potential field responses. The gravity field is computed assuming that each elementary cell mass is reduced to a single point at its center. The corrections for topography, free air and terrain effects leading to the Bouguer anomaly are considered in the calculation. The magnetic field components are computed assuming that the elementary cell is reduced to a magnetic dipole. Consequently, the computations can be used as the first approximation for small tetrahedra for which the distance between the observed point at surface and the tetrahedron center is much larger than the cell dimensions. In our development, we refined the near surface structure to subdivide recursively the coarse cells into 8 tetrahedra. This process is performed automatically and recursively for each tetrahedron as long as the previous condition is not respected. This technique is not only for near surface structure, it is also for anomalous zone such as faults, mineralization favorable zone etc., where there is possibility of sudden variation of physical property and thus of structure. The program has been tested on basic models (such as sphere, prism, etc.) with homogeneous parameters (density and susceptibility); the results show an excellent consistency with the analytical solution. Then, this program was applied to more complex models with heterogeneous parameters. As compared to equivalent Cartesian grid representation, the response accuracy is mainly the same but in the proposed program the computational time is better and faster than in the Cartesian grid representation.