Determining reservoir geometry and properties from the surface is a challenging but important task. The seismic method is in most situations the geophysical tool of choice, providing high resolution images of structure and stratigraphy. Given high quality seismic data, and well log information for calibration, seismic data can be inverted to provide quantitative information on reservoir properties such as porosity and lithology. However, in many situations when taken alone seismic can struggle to discriminate fluids and their saturations.
Electrical resistivity is well known to respond strongly to changing fluid type and saturation, which is why the resistivity log is such an important component of any well log suite, and is an integral part of a petrophysical interpretation. The Controlled Source ElectroMagnetic (CSEM) method allows resistivity to be measured from the seafloor (Figure 1). Taken by itself, however, this measurement of resistivity can also be ambiguous. The structural resolution of the method is poor, and without further constraint the depth of features of interest may be uncertain. Nor can regions of high resistivity be uniquely linked to subseafloor hydrocarbon deposits: they could equally be caused by low porosity sands or by carbonates, salt bodies or volcanics, among other things.
However, when CSEM is combined with seismic reservoir characterisation approaches within a rock physics framework, many of these ambiguities can be resolved, and the complementary measurement of resistivity derived from CSEM can significantly improve the accuracy with which reservoir properties in general, and saturation in particular, can be determined.
Although there are many benefits to putting seismic, CSEM and well log datasets together, there are also a number of challenges.
The first challenge is one of physics. Seismic and CSEM methods probe the earth using very different physical processes. The resulting electric and elastic properties must be linked in a consistent fashion to the underlying rock and fluid properties. This requires a rock physics framework to be either numerically derived or empirically calibrated at well locations. In either case the models are subject to uncertainty, which in turn leads to uncertainty in the resulting interpretation.
The second challenge is one of scale. Seismic, CSEM and well log techniques sample the earth at very different scales, varying from a few centimetres in the case of well logs, to hundreds of metres for CSEM. For example, using seismic data it may be possible to resolve the individual sands within a stacked reservoir interval. However, CSEM methods are likely only to resolve the bulk properties of the interval and not the details of the fluid distribution in the individual sands within it. These different scales must be reconciled in an integrated interpretation or joint inversion approach. Again, a robust rock physics framework, calibrated to well logs, is required to reduce this uncertainty.
The final challenge is one of sensitivity. It is perhaps an obvious statement, but in order for an integration approach to be successful, both seismic and CSEM methods must be sensitive to the interval of interest and changes in properties within it. This is a key consideration in determining where multi-physics approaches can be applied. For example, a reservoir that can be imaged and constrained seismically may lie at too great a depth below mudline, or be embedded in too complex or resistive a background structure for CSEM methods to be effective. Similarly, low saturation gas clouds above a reservoir may render seismic ineffective, whilst having little or no effect on the CSEM response or interpretation. Establishing the common domain of sensitivity is a key first step in any multi-physics project.
What is Meant by Integration?
The terms ‘multi-physics analysis’ or ‘integrated interpretation’ can be used to cover a huge variety of different approaches and workflows, ranging from the purely qualitative to the quantitative.
Perhaps the simplest approach to combining CSEM and seismic data is co-rendering CSEM inversion results with seismic sections (in the depth domain). The correlation (or lack thereof) between seismic structures and CSEM-derived resistivity variations can be used to draw inferences on the geology under study. This is a powerful first-look tool, but must be treated with care. Resistivity sections from unconstrained inversions suffer from the inherent low vertical resolution of the CSEM method when interpreted alone, and the depth to features of interest may be uncertain or incorrect. If co-rendered and interpreted without care alongside seismic this could lead to erroneous geological conclusions. This limitation can to some extent be mitigated by using seismic-derived structural constraints in the inversion of CSEM data to improve the vertical resolution; however, care must still be exercised. Perhaps the most important pitfall is in interpreting the cause of resistivity variations observed. A zone of high resistivity could as easily be caused by a lithological variation as by a change in hydrocarbon charge. Correlation between structural closures or fluvial stratigraphy and zones of high resistivity can perhaps provide confidence in an interpretation of hydrocarbon presence, but this is not infallible and the underlying question remains: what is the cause of the resistivity variation?
Integrated interpretation approaches start to address this question, by using rock physics relationships to quantitatively interpret the observed variations in seismic and CSEM-derived attributes in terms of the underlying rock and fluid properties. This is a two-stage process, and is illustrated in Figure 2. First, seismic data are inverted to provide acoustic and/or elastic impedance, and derived properties such as Poisson’s ratio. Similarly CSEM data are inverted, usually with structural constraints from seismic included, to provide a measure of resistivity and resistivity anisotropy. Second, these physical properties are then coupled through rock physics relations and jointly interpreted to provide an understanding of variations in rock and fluid properties.
Joint inversion approaches seek to do both steps at once by inverting directly for a model that satisfies simultaneously both seismic and CSEM datasets. A number of approaches to this problem have been suggested, all of which require electric and elastic domains to be coupled. There are broadly two approaches to achieving this coupling. In the first category are methods where the coupling between electric and elastic properties is primarily structural, constraining variations in resistivity to be spatially coincident with variations in elastic properties. Such methods are useful in cases where the variations in properties are known to be coincident, or where a direct relationship between electric and elastic properties is uncertain or difficult to obtain. An alternative approach uses rock physics models to relate the elastic and electric properties of the earth directly to underlying lithology and fluid properties. This is a good approach in many situations, but uncertainties in these rock physics relationships and the possibility that they vary away from well control (for example with lateral lithology variations) can increase the uncertainty in the inversion result.
The Fluid Content of Chalk
An example of the benefits of an integrated interpretation approach is shown in Figure 3. In this example the challenge is to determine the fluid content of a chalk reservoir in the North Sea. In this case seismic data can be used to accurately map the porosity variations in the chalk, but when taken alone have little sensitivity to the fluid content. In contrast CSEM data provided a measurement of resistivity within the chalk, and indeed identified a high resistivity zone. The cause of this is, however, uncertain when only resistivity is considered: is it the result of hydrocarbon fluids saturating an area of high porosity, or simply an area of low porosity chalk?
Taking the two measurements together resolves this ambiguity. Once the porosity variations in the chalk are constrained with seismic data and accounted for in the resistivity map, any remaining resistivity variations have a high probability of being related to changes in fluid properties, and in this case a confirmed hydrocarbon reservoir. Linking the seismic and electric domains through well log calibrated rock physics models further allows quantification of reservoir properties, in this case hydrocarbon column height.
Benefits of a Multi-Physics Approach
The demands of reservoir characterisation often go well beyond a simple requirement for an image of structure, impedance or resistivity. In order to fully understand the characteristics of a field, quantitative estimates of properties such as porosity, clay content, fluid type and saturation are required. However, such estimates can be fraught with uncertainty when only a single data type is considered. Seismic data remains the backbone of geophysical analysis and reservoir characterisation, but the advent of CSEM methods, measuring seafloor resistivity, brings the possibility of integrating multi-physics data within the characterisation workflow. Such approaches are of course not infallible, nor are they universally applicable. They must also be tailored to the problem of interest. However, when carefully applied, integrated interpretation approaches have the potential to significantly improve the robustness of reservoir characterisation projects.