Improved reservoir management and production optimisation demands require accurate characterisation of reservoir properties and their changes through time.  Advances in geophysical data acquisition and interpretation have led to significant improvements in the remote imaging of earth structure and properties.  However, when only a single data type is considered, ambiguities in the interpretation can remain. Integration of disparate geophysical data types allows the strengths of each to be exploited.  Here we will concentrate on three contrasting methods: surface seismic, marine controlled source electromagnetic (CSEM) and well-log data. 

Well logs provide a high resolution measurement of the properties of a reservoir and the surrounding strata, however properties can only be determined in a small area local to the well.  Often a measurement of reservoir properties across the extent of a field are desirable for reservoir management or production optimization.  Remote geophysical measurements are therefore required.

Seismic data are commonly used to provide images of sub-surface structure, from which high resolution geological models of structure and stratigraphy can be developed.  Amplitude variation with offset (AVO) and inversion for acoustic and elastic impedance may also be used to constrain properties such as elastic moduli, porosity and density. However seismic data alone in many situations cannot give a complete picture of the reservoir.  For example, AVO anomalies may be caused either by fluid or lithological variations, which cannot be separated on the basis of the seismic data alone.

CSEM methods use a high powered source to transmit low frequency signals through the earth to an array of receivers.  By interpreting the received signals using forward modelling and inversion approaches, the resistivity structure of the seafloor can be determined. In many situations electrical resistivity is driven by the properties and distribution of fluids in the earth. Resistivity well logs often show that commercial hydrocarbon deposits may be many times more resistive than surrounding lithologies.  In principal such variations should be readily detected using CSEM tools. In contrast, seismic data are sensitive to boundaries between lithologic units but are less sensitive to fluid changes within these units. Given high quality seismic and well data and sophisticated seismic inversion and rock physics tools, we can sometimes relate these seismic changes to saturation effects. Nevertheless, the change in resistivity caused by variations in saturation should be much easier to detect. 

However, despite the increased sensitivity of resistivity data over seismic for the determination of saturation, there are two inherent challenges to interpreting CSEM data. Firstly, the structural resolution of CSEM data is poor.  Secondly, the cause of resistivity anomalies (particularly high resistivity features) cannot be uniquely linked to the presence of hydrocarbons in the subsurface when taken in isolation.  In many situations these are equally likely to be caused by other high resistivity material (for example, tight carbonates, salt or volcanics). Both of these limitations must be addressed when considering the applicability of CSEM to answer a geophysical question, and as far as possible mitigated by the interpretation approach adopted. 

CSEM data can, of course, be interpreted in isolation, and if there were no seismic data or wells in the vicinity of the CSEM dataset (for example if a survey were performed in a frontier area), then this would be necessary. However, with no constraints on this interpretation, the result will suffer from the non-uniqueness and ambiguity which blight unconstrained interpretation approaches.  Although resistivity is imaged, the poor structural resolution of the method means that such images are diffuse and difficult to interpret.   The uncertainty in the depth of features is large, so that they cannot be unambiguously attributed to a particular stratum.  If there are multiple resistive features, these cannot be easily separated, and small resistive bodies are likely to be lost or smoothed into surrounding strata.    Even assuming that localized resistivity anomalies can be found, the cause of these anomalies cannot be unambiguously linked to the presence of hydrocarbon.  

In the presence of seismic and well information, the question that we are trying to answer with the CSEM data becomes significantly better posed. The question is no longer one addressed at finding a reservoir, but rather one of determining the content of a defined structure. Using seismic information the reservoir structure is known (but potentially not its content or extent), and we have independent constraints on the surrounding strata within which it is embedded. This is therefore a constrained interpretation problem and one that the CSEM data are in a much better position to answer. 

It is clear that a careful combination of all three data types can supply information that is not available, or is unreliable from any one data type alone. By integrating complementary sources of information and exploiting the strengths of each, estimates of rock and fluid properties such as gas saturation and porosity can be obtained with greater confidence than from any one data type alone.