GB2458959A - Electromagnetic surveying using a crossed electric dipole transmitter with selectable orientations - Google Patents

Abstract

Various controlled source electromagnetic survey methods for surveying subterranean strata are described. The methods rely on an electromagnetic source 22 comprising two orthogonal horizontal electric dipole transmitters Tx, Ty. Broadcast signals from the two transmitters combine to provide a corresponding resultant virtual dipole transmitter v. The effective orientation of the resultant dipole transmitter v may be arbitrarily selected so that data for arbitrary transmitter orientations may be provided (e.g. inline or broadside as desired). This is achieved without needing to move a source around, or to provide separate sources for each desired orientation. In some examples, a desired resultant virtual transmitter orientation may be provided by applying appropriately scaled drive signals Sx, Sy to the two transmitters. In other cases, a desired effective resultant virtual transmitter orientation may be provided by applying equal drive signals to the two transmitters, but scaling the receiver signals associated with each transmitter relative instead. I.e. in some cases the relative scaling may be applied at the transmission stage, and in other cases the relative scaling may be applied at the detection stage.

Description

TITLE OF THE INVENTION

ELECTROMAGNETIC SURVEYING

BACKGROUND ART

The invention relates to seafloor electromagnetic surveying for resistive and/or conductive bodies, for example for oil, gas, methane hydrates etc. and other hydrocarbon reserves or subterranean salt bodies.

Seismic techniques are frequently used during hydrocarbon-exploration expeditions to identit' the existence, location and extent of reservoirs in subterranean rock strata. However, whilst seismic surveying is able to identify such structures, the technique is often unable to distinguish between the different possible compositions of pore fluids within them. This is especially so for pore fluids which have similar mechanical properties, such as oil and seawater. It is therefore generally necessary to employ other techniques to determine whether a reservoir contains oil, for example, or just aqueous pore fluids. One group of such techniques are the controlled source electromagnetic (CSEM) surveying techniques.

CSEM surveying techniques seek to distinguish oil-and water-filled reservoirs on the basis of their differing electrical properties. The techniques typically involve transmitting an electromagnetic (EM) signal into the seafloor, generally using a horizontal electric dipole (FLED) source / transmitter, and measuring the response at EM receivers /detectors for a,range of distances from the source (i.e. for a range of source-receiver offsets / separations) [1, 2, 3, 4]. Resistivity maps of subterranean strata, perhaps to depths of several kilometres, may be recovered from the measured detector responses, e.g., using conventional forward modeling, geophysical inversion and/or imaging/migration techniques.

In CSEM surveying it is important to consider the orientation of current flows induced in subterranean strata by a transmitted EM signal (I]. This is because the response of subterranean strata (which in many situations will comprise planar layers) is generally different for transverse magnetic (TM) mode components of a transmitted

I

signal, which excite significant components of vertical current flow, and transverse electric (TE) mode components, which excite predominantly horizontal current flows.

For TE mode components, the coupling between the generally horizontal layers comprising the subterranean strata is largely inductive. This means the presence of a s thin layer of contrasting resistivity (e.g. such as a higher resistivity hydrocarbon reservoirs) does not significantly affect EM fields detected at the surface. This is because the large scale TE mode current flow pattern is not greatly affected by a thin layer of contrasting resistivity.

However, for TM mode components, the coupling between layers includes a significant galvanic component (i.e. transfer of charge between vertically separated layers). This means, for example, even a thin resistive layer can strongly affect the TM mode components of EM fields detected at a receiver. This is because the large scale TM mode current flow pattern is interrupted by the thin resistive layer. It is thus known that the TM mode of coupling is useful for constraining thin planar layers of contrasting resistivity, such as a hydrocarbon reservoir.

Accordingly, there is often a desire in CSEM surveying to provide data which are primarily sensitive to the TM mode of coupling because this mode is more sensitive to the presence of thin layers of contrasting, e.g. higher, resistivity than the TE mode.

However, in some cases sole reliance on TM mode components can lead to ambiguities in the interpretation of the data. This is because as well as being sensitive to thin layers, TM mode data are, like TE mode data, also sensitive to the larger scale (background) resistivity structure. This means, for example, the effects of a thin resistive layer on detected fields can be indistinguishable from the effects of other realistic large scale subterranean strata configurations. Because of this potential ambiguity, it is known that collecting data which are primarily sensitive to thin resistive/conductive layers (i.e. TM mode dominated) and data which are primarily sensitive to the large-scale background structure only (i.e. TE mode dominated) can in many situations provide more for a comprehensive survey [I].

Various different techniques have been proposed to provide data that are differently sensitive to the TM and TE modes of coupling) [1, 2, 3, 4]. Sometimes only TM mode data will be desired for a given survey (e.g. because the large-scale background structure is already known so there is reduced risk of ambiguity in interpretation). In other cases both TM and TE data may be desired.

One known way of obtaining respective TM-and TE-mode dominated data sets is to obtain data for different relative orientations between an MED transmitter and receivers in an array [1). According to this scheme, data from receivers which are arranged inline with the MED source (i.e. on a line parallel to and passing through the dipole axis of the MED transmitter) are TM mode dominated (sensitive to the presence of thin resistive layers indicative of hydrocarbon-bearing reservoirs), whereas data from receivers which are arranged broadside to the I-LED source (i.e., on a line perpendicular to and passing through the HED axis), are TE mode dominated (more sensitive to characteristics of the large scale background). This approach may be referred to as a geometric splitting approach. The splitting arises from the inherent geometry of the dipole field generated by the source. During analysis, the complementary data sets (i.e. the respective inline and broadside data sets) may be combined to reveal differences between the TE mode and TM mode coupling between the transmitter and the detector. These differences can help to resolve the above-mentioned potential for ambiguity in interpretation of CSEM data. (The principle of geometric splitting may also be applied to surveys in which only TM-dominated data are required (e.g. in cases where the large scale background structure is already known), in which case primarily inline data may be obtained. Similarly, the principle may be applied to surveys in which only TE-dominated data are required (e.g. to map the large-scale background resistivity structure), in which case primarily broadside data may be obtained.) Figure 1 shows in plan view an example schematic survey geometry for collecting respective TM and TE dominated data sets according to the geometric splitting approach. There are sixteen receivers 25, and these are laid out in a square grid on a section of seafloor 6 above a subterranean hydrocarbon reservoir 56. The hydrocarbon reservoir 56 has a boundary indicated by a heavy line 58. The orientation of the hydrocarbon reservoir is indicated by the cardinal compass points (marked N, E, S and W for North, East, South and West respectively) indicated in the upper right of the figure. To obtain respective TM and TE mode data sets, a source such as a HED antenna, starts from location A' and is towed along a path indicated by the broken line

S

through location B', until it reaches location C,, which marks the end of the survey path. As is evident, the tow path first covers four parallel paths aligned North-South. This portion of the survey path moves from location A' to location B'.

Starting from location B', the survey path then covers four paths aligned East-West.

Each detector is thus passed over in two orthogonal directions. The survey is completed when the source reaches the location marked C'.

During the towing process, each of the receivers 25 presents several different orientation geometries with respect to the source. For example, when the source is directly above the detector position DI and on the North-South aligned section of the tow path, the detectors at positions D5, D6 and D7 are at different ranges in an end-on position (and thus providing data dominated by TM mode coupling), the detectors at positions D2, D3 and 04 are at different ranges in a broadside position (and thus dominated by TE mode coupling), and the detector at positions D8 and D9 are midway between (providing mixed TM and TE mode data). When the source later passes over the detector position Dl when on the East-West aligned section of the tow path, the detectors at positions 05, D6 and D7 are now in a broadside position, and the detectors at position D2, D3 and D4 are in an end-on position. Thus, in the course of a survey, and in conjunction with the positional and orientation information of the source, data from the detectors can be used to provide details of the source electromagnetic signal transmission through the subterranean strata for a range of distances and orientations between source and detector. Each orientation provides varying TM and TE mode contributions to the signal propagation. Thus TM and TE mode dominated data sets may be provided by appropriate collation of data from the different receivers at different times.

A drawback of the geometric splitting approach, such as shown in Figure 1, is the need to obtain data for specific source-receiver alignments results in complex and lengthy tow paths. These are costly and time consuming to perform, and furthermore provide relatively little data that is significantly dominated by one or other of the TM and TE modes of coupling. This is because for the majority of the time the individual receivers are neither inline nor broadside with respect to the dipole source, but are at an intermediate position. Similar comments apply even if primarily TM (inline) only or TE (broadside) only data are from a two-dimensional survey area. This is because

S

lengthy tow paths are still required to obtain data from a 2D array of receivers for the desired transmitter orientation.

The geometric splitting approach may also be applied to a fixed installation (i.e. non-towed) application of CSEM. However, in this case large numbers of differently located and orientated transmitters and receivers are required to provide appropriate spatial coverage. For example, a fixed installation survey for providing comparable spatial coverage of TM and TE dominated data to that of Figure 1 would in effect need transmitters to be located at numerous locations along the tow path in that figure.

One proposed way of obtaining TM and/or TE dominated data sets which does not rely on geometric splitting is to mathematically decompose CSEM data into TM and TE mode components (2, 3]. This can be done for data obtained over a range of source-receiver orientations and so more efficient tow paths can be employed compared to schemes based on geometric-splitting. However, these mathematical decomposition techniques are relatively sensitive to noise, and also require specialised receiver (and/or transmitter) designs capable of measuring spatial gradients in

electromagnetic field components.

WO 2006/059122 [7] recognizes that an HED transmitter provides for geometric splitting and considers this to be disadvantageous. The aim of WO 2006/059 122 is thus to avoid the geometric splitting by providing a source which emits TE and TM modes of approximately equal amplitudes in all directions simultaneously. To achieve this, the source in WO 2006/059 122 comprises multiple electrodes with non-coincident time-varying signals being applied to different pairs of the electrodes to produce a rotating electric fieLd. The rotating field is said to provide for simultaneous TE and TM mode transmission in all direction. This approach thus means all receivers obtain corresponding combined TE and TM data regardless of their location. Thus the geometric splitting associated with a linear dipole source is avoided and simpler tow paths may be used to provide corresponding data to a range of receivers over a 2D array. It is not clear from WO 2006/059122 how the combined TE and TM mode signal may be separated at the receivers so that distinct TE and TM mode data are obtained.

I

WO 2004/053528 [8] discloses an application of CSEM surveying to hydrocarbon reservoir monitoring. An array of transmitter electrodes are provided over an area of interest. Each electrode pairing may be considered to provide for a different horizontal electric dipole transmitter. Thus as many different horizontal electric dipole transmitters are provided as their are different pairing of transmitter electrodes (so long as appropriate drive signal switching / routing is provided). However, this approach still only provides for a finite range of horizontal electric dipole transmitters.

In view of the above, there is a need for CSEM survey techniques which allow data for different modes of coupling (e.g. TM and TE, or a selected mix of TM and TE) to be collected as desired for a given implementation and which do not suffer from the above-described drawbacks of known techniques.

S

SUMMARY OF TILE INVENTION

According to a first aspect of the invention there is provided a controlled source electromagnetic survey method for surveying an area thought or known to contain a subterranean resistive or conductive body, the method comprising: providing a first dipole transmitter aligned along a first direction; providing a second dipole transmitters aligned along a second direction; applying a first drive signal to the first transmitter so as to broadcast a first source signal; and applying a second drive signal to the second transmitter so as to broadcast a second source signal, wherein the relative strengths of the first and second drive signals are selected according to a desired orientation of a resultant source signal corresponding to a combination of the first and second source signals.

Thus a resultant source signal corresponding to an arbitrarily oriented virtual horizontal electric dipole transmitter may be provided by applying appropriately scaled drive signals to the first and second transmitters. This may be referred to as source-based scaling. Data obtained at receivers for the virtual dipole transmitter may processed and interpreted I analysed in accordance with conventional techniques as if the virtual transmitter were a real transmitter.

The first and second directions may be orthogonal. This is not necessary, but provides for a simpler determination of the appropriate relative scaling for the strengths of the two drive signals. A ratio of the strengths of the first and second source signals may, for example, be approximately equal to a tangent of an angle between the desired orientation of the resultant source signal and one of the first and second directions.

The first and second source signals may be broadcast simultaneously, e.g. in phase, or at different times. When broadcast at different times, receiver data recorded in response to the respective first and second source signals and seen at corresponding times within the respective broadcast periods may be summed to provide combined receiver data corresponding to that which would be seen were if the first and second source signals were broadcast simultaneously.

The method may further comprise providing a receiver for detection of source signals broadcast from the first and second transmitters, wherein the desired orientation of the resultant source signal is selected according to an angular position of the receiver with respect to one of the first and second directions. For example the desired orientation of the resultant source signal may be an inline orientation with respect to the receiver.

The desired orientation of the resultant source signal may also be in a broadside orientation with respect to the receiver. Any other intermediate orientation could also be provided. This means commonly used orientations for CSEM surveying, e.g. inline and broadside orientations, can be provided regardless of a receiver's angular position with respect to the transmitters.

The relative strengths of the first and second drive signals may be changed during the survey to cause a change in orientation of the resultant source signal corresponding to the combination of the first and second source signals (the resultant source signal maintains a generally linear dipole form). Thus a virtual I-LED transmitter may be provided that may be sequentially oriented as desired (e.g. inline, broadside or in between) with respect to receivers over a range of angles. Thus a virtual HED transmitter may be provided that in effect scans over a survey area / jumps between different orientations, so as to provide data for a desired orientation for a range of receivers at is different azimuths.

According to a second aspect of the invention there is provided a controlled source electromagnetic survey method for surveying an area thought or known to contain a subterranean resistive or conductive body, the method comprising: providing a first dipole transmitter aligned along a first direction; providing a second dipole transmitters aligned along a second direction; obtaining a first data set at a receiver for a first source signal broadcast by the first transmitter; obtaining a second data set at the receiver for a second source signal broadcast by the second transmitter; scaling one of the data sets relative to the other by a relative scaling factor determined according to a desired orientation of an effective resultant source signal corresponding to a combination of the first and second source signals; and combining the relatively scaled first and second data sets to provided a combined data set.

The first and second aspects of the invention are complementary aspects in that they rely on broadly the same principles. However, in accordance with the first aspect of the invention a relative scaling is applied to drive signals at the transmission stage, whereas in accordance with the second aspect of the invention, an effective relative scaling of the drive signals is provided for by scaling receiver signals associated with each transmitter after detection. The approach of the second aspect of the invention may thus be referred to as a receiver-based scaling approach.

A further advantage of the receiver-based scaling approaches, i.e. approaches based on scaling the strength of the signals comprising at least one of the data sets s (e.g., scaling-up or -down the amplitude / intensity of recorded electric and I or magnetic fields), is that once the two data sets have been obtained, they may be scaled for any desired effective virtual dipole source orientation without requiring any further source transmissions. This means data suitable for receiver-based scaling can be differently scaled and combined as desired to provide any number of corresponding virtual dipole transmitter orientations, even after the survey has finished. Another significant advantage of receiver-based scaling approaches is that the drive signals applied to the transmitter are the same regardless of any final desired orientation (i.e. the drive signals are not tailored to take account of any specific receiver's azimuth).

This means multiple receivers at different azimuths can collect their own first and second data sets simultaneously. The data sets for each individual receiver may then be scaled and combined as appropriate to provide data corresponding to any desired virtual dipole orientation for any of the receivers.

As with the first aspect of the invention, the first and second directions may be orthogonal.

The first and second source signals may be broadcast at different times. This provides a ready way of distinguishing data from the two transmitters so that the two data sets may be separately obtained. In this case the first and second source signals may be broadcast in a series of alternating short broadcast periods. This can help minimise the impact of any time dependence on the measured data (e.g., because the source is moving, or because of potential instrumental drifts.) In cases where the first and second source signals are broadcast at different times, there are periods when the first transmitter is not transmitting the first signal, and periods when the second transmitter is not transmitting the second signal. To maximisethe useful duty cycle of the respective transmitters, the first and second source signals may primarily comprise components at a first frequency, and the method may further comprise obtaining a third data set at the receiver for a third source signal broadcast by the first transmitter at times when the second transmitter is broadcasting the second signal, and obtaining a fourth data set at the receiver for a fourth source signal broadcast by the second transmitter at times when the first transmitter is broadcasting the first signal, and wherein the third and fourth source signals primarily comprise components at a second frequency different from the first frequency.

This approach thus allows virtual dipole transmitters of arbitrary orientation to be provided for using receiver-based scaling at both the first frequency and the second frequency. This approach thus not only makes more efficient use of the available transmitter time for each transmitter, it also increases the bandwidth of data obtained in the survey.

In some embodiments according to the second aspect of the invention, the first and second source signals may be broadcast simultaneously. In this case the first and second source signals may comprise different frequency components. The differing frequency content provides another ready way of distinguishing data from the two transmitters so that the two data sets may readily be separately obtained. Furthermore, the first source signal may comprise a first fundamental frequency component and the second source signal may comprise a second fundamental frequency component, and the second fundamental frequency component may be at a frequency that is an even multiple of the frequency of the first fundamental frequency component This can allow inter-frequency interpolation, for example by interpolating between data at two or more frequency components in one of the data sets to provide interpolated data at a frequency corresponding to a frequency component in the other data set. This can aid interpretation as data can be provided for both transmitters at what is in effect the same frequency.

The relative scaling factor may be approximately equal to a tangent of an angle between the desired orientation of the effective resultant source signal and one of the first and second directions.

Furthermore, the desired orientation of the effective resultant source signal may be selected according to an angular position of the receiver with respect to one of the first and second directions. For example, the desired orientation of the effective resultant source signal may be an inline or a broadside orientation with respect to the receiver.

According to a third aspect of the invention there is provided a method of monitoring a reservoir, comprising performing a survey of the area known to contain the reservoir according to the first or second aspect of the invention at a first time, -Il-performing a survey of the area known to contain the reservoir according to the first or second aspect of the invention at a second time, and comparing results from the survey performed at the first time with results from the survey performed at the second time to identify changes in the reservoir between the first and second times.

According to a fourth aspect of the invention there is provided a computer readable storage medium having electromagnetic field data obtained during surveying according to the first or second aspect of the invention recorded thereon.

According to a fifth aspect of the invention there is provided a method for obtaining hydrocarbon from an area that contains a subterranean hydrocarbon reservoir, comprising: penetrating the subterranean hydrocarbon reservoir with a hydrocarbon-producing well; extracting hydrocarbon from the subterranean hydrocarbon reservoir using the hydrocarbon-producing well; surveying the reservoir using a method according to the first or second aspect of the invention in order to monitor depletion of the reservoir; and continuing to extract hydrocarbon from the hydrocarbon-producing well.

According to a sixth aspect of the invention there is provided a volume of hydrocarbon extracted according to the method the fifth aspect of the invention.

According to a seventh aspect of the invention there is provided a method of processing results from a controlled source electromagnetic survey of an area thought or known to contain a subterranean resistive or conductive body, the method comprising: providing a first data set obtained at a receiver for a first source signal broadcast by a first dipole transmitter aligned along a first direction; providing a second data set obtained at the receiver for a second source signal broadcast by a second dipole transmitter aligned along a second direction; scaling one of the data sets relative to the other by a relative scaling factor determined according to a desired orientation of an effective resultant source signal corresponding to a combination of the first and second source signals; and combining the relatively scaled first and second data sets to provided a combined data set.

According to a eighth aspect of the invention, there is provided a source for controlled source electromagnetic surveying, comprising: a first dipole transmitter aligned along a first direction; a second dipole transmitter aligned along a second direction; and a source control unit operable to provide first and second drive signals having selectable relative strengths to the respective first and second transmitters.

The source control unit may further be operable to determine relative strengths for the first and second drive signals according to a desired orientation of a resultant source signal corresponding to a combination of first and second source signals broadcast by the first and second transmitters in response to their respective drive signals.

Sources according to the eighth aspect of the invention may be used for surveying in accordance with the first or second aspect of the invention. f -13-

BRIEF DESCRIPTION OF THE DRAWING

For a better understanding of the invention and to show how the same may be carried into effect reference is now made by way of example to the accompanying drawings in which: Figure 1 schematically represents in plan view an electromagnetic survey according to the known techniques; Figure 2 schematically represents in plan view an electromagnetic survey according to an embodiment of the invention; Figure 3 schematically shows an xyz-Cartesian coordinate system and a cylindrical coordinate system used to describe some embodiments of the inventions; Figure 4A schematically shows the source and one of the receivers shown in Figure 2 in more detail during in one operating mode; Figure 4B schematically shows the source and one of the receivers shown in Figure 2 in more detail during in another operating mode; Figure 5 schematically represents in plan view an electromagnetic survey according to another embodiment of the invention; Figure 6A and 6B schematically show the source and one of the receivers shown in Figure 2 in more detail during other operating modes; Figure 7 schematically represents in plan view an electromagnetic survey according to another embodiment of the invention; Figure 8 schematically represents in vertical section view an electromagnetic survey according to an embodiment of the invention using a towed source; Figure 9 schematically shows a towable source for use in the survey represented in Figure 8; Figure 10 schematically represents in plan view the electromagnetic survey represented in Figure 8; Figures 11 and 12 schematically show different example drive signals for applying to a source during surveying in accordance with embodiments of the invention; Figure 13 is a schematic view of an oil rig producing hydrocarbon according to an embodiment of the invention; Figure 14 is a schematic perspective view of a barrel containing a volume of hydrocarbon according to an embodiment of the invention; and Figure 15 is a schematic perspective view of a data storage medium bearing a data set according to an embodiment of the invention.

-15 -

DETAILED DESCRIPTION

Figure 2 schematically shows in plan view an electromagnetic survey according to an embodiment of the invention. The survey is of an area that is similar to that shown in Figure 1 in that it comprises a section of seafloor 6 above a subterranean hydrocarbon reservoir 56 having a boundary 58. The orientation of the hydrocarbon reservoir is indicated by the cardinal compass points (marked N, E, S and W for North, East, South and West respectively) indicated in the upper-right of the figure.

An electromagnetic source 22 is located near to the centre of the reservoir. The source 22 comprises an orthogonal pair of horizontal electric dipole (HED) transmitters. A first HED transmitter T is aligned parallel to an East-West line and a second 1-LED transmitter T is aligned parallel to a North-South line. These specific alignments are assumed here purely for ease of explanation; in a practical survey the source may be arbitrarily aligned with respect to these axes. The individual I-LED transmitters comprising the new source 22 may be similar to HED transmitters of the type conventionally used in CSEM surveying.

The directions of x-, y-and z-axes of a Cartesian coordinate system used for describing some embodiments of the invention is also schematically shown in Figure 2 to the top-left. The x-axis runs parallel to the dipole axis of the first transmitter T, and the y-axis runs parallel to the dipole axis of the second transmitter Is,. Thus the first transmitter may be referred to as the x-transmitter and the second transmitter may be referred to as the y-transmitter. The z-axis runs vertically downwards into the seafloor.

Although the Cartesian coordinate system is shown displaced from the source 22 in Figure 2 for clarity, in general the origin of the Cartesian coordinate system will be considered to be at the centre of the source 22 (i.e. where the two transmitters cross in this example configuration).

The source 22 further includes a drive unit (not shown). The drive Unit IS operable to supply respective first and second drive signals S,, S, to the first and second HED transmitters T, I. (The first drive signal may be referred to as the x-drive signal and the second drive signal may be referred to as the y-drive signal). The form of the drive signals may be any waveform suitable for CSEM surveying. For -16-example, a 0.25 Hz square-wave might be used. Other frequencies and waveform shapes could also be used as is conventional.

The x-transmitter I comprises a pair of electrodes separated along the x-axis, in this example by 300m, and arranged to receive the x-drive signal S from the drive unit. The y-transmitter I, comprises a pair of electrodes separated along the y-axis, again by 300rri in this example, and arranged to receive the y-drive signal S from the drive unit. When the HED transmitters are supplied with their respective drive signals they broadcast EM signals into the seawater. The amplitude of the drive signal applied to a transmitter of a given length will determine the maximum dipole moment of the transmitter for the applied waveform, and so for simplicity will be parameterised as such. Thus a drive signal S, applied to transmitter T will be taken to mean that transmitter T has a maximum dipole moment S for the waveform applied (i.e. a maximum current 5,1300 is applied (because it is a 300m long dipole)). The drive unit in this example comprises two broadly conventional CSEM cycloconverters (e.g., such as describe in GB 2415 785 [5]) operating in parallel. The cycloconverters (or at least one of them) are configured so that the relative strengths of the drive signals supplied to the two I-LED transmitters may be varied (e.g. by changing the relative electric currents supplied to each transmitter). Drive units based on other conventional CSEM signal generation technologies, e.g. based on switched DC sources such as described in GB 2415 785 [9], could equally be used.

The source 22 in this embodiment is taken to remain stationary throughout a survey (in other examples, as discussed further below, a source may be continuously towed during surveying). A fixed installation may be useful, for example, for monitoring applications. For example, the method could be used to monitor changes to a hydrocarbon reservoir from which hydrocarbon is being drawn. Such monitoring can allow the extraction process to be optimized according to changes in the reservoir. The HED transmitter(s) comprising the source could thus be anchored to the seafloor or suspended from an oil-rig platform, for example. In other examples, the RED transmitter(s) could be placed in a horizontal well or borehole, e.g. a geotechnical borehole. In still other cases the source 22 may simply rest on seabed, e.g. having being lain there by a tow vessel that waits as data is collected before towing the source 22 to another location. a

An array of receivers 25, in this case an array comprising sixteen receivers, is laid out in a grid located generally over the reservoir 56. The receivers may be conventional receivers of the type normally used in CSEM surveying. In this example the receivers are remotely deployed receivers, e.g. similar to those described in WO s 03/104844 [61. Thus each receiver 25 includes an instrument package, a detector antenna, a floatation device and a ballast weight. The detector antenna comprises three orthogonal electric dipole antennae (generally two horizontal and one vertical, but other configurations may be used), and three orthogonal magnetic dipole antennae (again generally two horizontal and one vertical, but possibly in other configurations).

In other examples, fewer components of EM fields at the receiver may be measured, e.g. vertical field components may not be desired. The electric dipole detector antennae are sensitive to the electric field components of the electromagnetic fields induced by the source 22 in the vicinity of the detector, and produce electric field detector signals there from. The magnetic dipole detector antennae are sensitive to magnetic field components of the electromagnetic fields induced by the source 22 in the vicinity of the detector, and produce magnetic field detector signals there from.

The use of remotely deployed receivers such as shown in Figure 2 is an efficient way of sampling over a large area without requiring extensive infrastructure. However, in other example, e.g. for surveying / monitoring a relative small area, and/or for a long term fixed installation, receivers cabled back to a central receiver control station (e.g. near the source) may be employed instead with the central receiver control station providing power and control signals to the receivers and receiving data there from for storage or onward transmission.

Although the receiver array shown in Figure 2 is based on a square grid pattern, it will be understood that a wide variety of receivers placements may be used, for example other high symmetry regular grids, such as triangular or rectangular, may be used. An irregular array of receivers having no high level of symmetry could also be used.

Figure 3 shows a relationship between the x-and y-axes of the above-mentioned xyz-Cartesian coordinate system (which are respectively aligned with the first (Ta) and second (Tv) HED transmitters) and a cylindrical coordinate system which may also be used to describe relative source and receivers locations. The cylindrical coordinate system is centred on the source 22 (i.e. at the crossing point of the two transmitters T, Ii,) and is aligned such that the location of a receiver may be defined relative to the source 22 by a range (separation) R, an azimuth angle p, and a depth z. Range R is a distance from the centre of the source 22 to the receiver.

s The depth z (axial) coordinate extends vertically downwards into the seafloor 6 (i.e. depth z is not shown in Figure 3 for the cylindrical coordinate system, but corresponds with the z-axis of the Cartesian coordinate system shown in Figure 2).

Azimuth angle ç is an angle between the y-axis (i.e. a line passing through and running parallel to the y-transmitter dipole axis) and an imaginary line running from the source 22 to the location, measured clockwise when viewed from above. This angle may thus be referred to the azimuth of the receiver with respect to the y-transmitter (the angle also corresponds to the orientation of the y-transmitter with respect to the location defined location). It is noted that a similar azimuth angle may also be defined with respect to the x-transmitter. Such an azimuth angle would differ from the azimuth angle measured with respect to the y-transmitter by 900 since the x-and y-transmitters in this example are orthogonal). Thus a receiver at an azimuth (PR of 00 (or 1800) is in an exactly inline (end-on) position with respect to the y-transmitter, and in an exactly broadside position with respect to the x-transmitter. A receiver at an azimuthal angle (PR of 90° (or 270°) is in an exactly inline (end-on) position with respect to the x-transmitter, and in an exactly broadside position with respect to the y-transmitter. A receiver at an intermediate azimuthal angle (PR (i.e. an azimuth away from p = 00, 90°, 180° or 270°) is neither inline nor broadside with respect to either transmitter.

Figure 4A schematically shows the source 22 and one of the receivers 25 shown in Figure 2 in more detail (not to scale). The receiver is at an azimuth angle (PR.

and a range R with respect to the source. (Throughout this description locations will be assumed to be in a horizontal plane containing the source -i.e. it will be assumed z = 0 unless the context demands, or it is specifically stated, otherwise).

In addition to schematically showing the respective x-and y-transmitters T, Ia,, Figure 4A also schematically shows the strengths of the respective drive signals S,, S, applied to the transmitters. In this example the two drive signals S,, S,, are applied simultaneously to their respective transmitters and in phase and having the same overall shape of waveform, but with different amplitudes (strengths). The strengths of the drive signals S,, S, are schematically represented in Figure 4A as arrows having lengths corresponding to their respective amplitudes and aligned with their respective s transmitters. It will be assumed here that the cycloconverters in the drive unit are each capable of supplying a maximum current amplitude m for the waveform used, and that the schematic graphic representations of the drive signals in Figure 4A are scaled such that a drive signal corresponding to m would have a length in Figure 4A matching that of its transmitter and correspond to a dipole moment amplitude of Sm.

The receiver 25 shown in Figure 4A is at an azimuth of around (PR = 30°. The drive signal strengths S, S, are determined in this example in dependence on this angle, i.e. in dependence on the orientation of the receiver location with respect to the HED transmitters comprising the source. In this case the x-drive signal S, which is applied to the x-transmitter is Smax. Sifl((PR), and the y-drive signal S, which is applied to the y-transmitter is Sm. COS(CPR). Thus for (PR 30°, drive signal S, is approximately 50% of Smax and drive signal S, is approximately 87% of Smax. That is to say, a scaling factor f = 0.5 is applied for the x-drive signal and a scaling factor f = 0.87 is applied for the y-drive signal.

The combined resultant of the x-transmitter broadcasting with a dipole moment signal S, ( Smax. Slfl((PR)) and the y-transmitter broadcasting with a dipole moment signal S, ( Smax. COS((PR)) is equivalent to a single "virtual" HED transmitter V shown in Figure 4A as a dotted line. The orientation (v of the virtual dipole transmitter with respect to the "real" transmitters T, T (i.e. the angle between the axis of the virtual dipole V and the azimuth cp = 0°), is equal to (PR (because, in accordance with conventional vector geometry, y = tan 1(sin((PR)/cos(çR)). That is to say, the resultant virtual dipole V in this example is arranged inline with the receiver 25 at an azimuth angle 30°. In practice the combined resultant of the two drive signals S,, S applied to the two transmitters T, T will not be exactly equivalent to the virtual dipole shown in Figure 4A. However, as discussed further below, the error is small and can be mitigated / accounted for if desired. The magnitude S., of the virtual dipole is equal to Smax (i.e. S.,2 = S,2 + S2).

I

Thus, by providing appropriately scaled drive signals to the two HED transmitters comprising the source 22, a resulting virtual dipole HED transmitter having any desired orientation can be provided. That is because the virtual dipole transmitter having the desired orientation may be considered as being decomposed into s the two "real" dipoles aligned with the x-and y-transmitters. Conventional principles of vector geometry may thus be used to determine appropriate relative signal strengths to apply to the x-and y-transmitters to provide a combined resultant EM signal that is (approximately) equivalent to the desired virtual dipole transmitter. The appropriate drive signals will depend on the desired orientation of the virtual dipole transmitter io with respect to the x-and y-transmitters. Thus, and as shown in Figure 4A, the signals could be scaled such that the combined signals from the x-and y-transmitters correspond to a virtual dipole transmitter in an inline orientation with respect to a receiver at any arbitrary azimuth. This means TM mode dominated data (i.e. effective inline data) can be obtained at a receiver regardless of its azimuthal orientation with respect to the individual HED transmitters T,, T comprising source 22 (i.e. without needing to move I rotate the source 22).

In the general case, for a source comprising two mutually orthogonal HED transmitters, a virtual dipole may be provided having any desired orientation py by applying signals to the two transmitters that are scaled with respect to one another such that S/S, = tan(p,); for example with S, = Smax. sin(pv) and Sy = Smax. cos((pv). It is noted that while in practice it may be simpler for the two HED transmitters comprising the source 22 to be orthogonal, they need not be. This is because any orientation of virtual dipole can still be provided even if the two dipole transmitters comprising the source are not orthogonal to one another (so long as they are not exactly parallel).

Figure 4B is similar to, and will be understood from Figure 4A, but shows a situation in which different scalings are applied to the respective transmitter drive signals to provide for a differently oriented virtual dipole V. For this example it is assumed that it is desired to provide a virtual dipole that is broadside with respect to the receiver 25 at azimuth angle pg = 30°. The desired orientation for the virtual dipole in this case is therefore Pv = pa + 90° (i.e. 120°). Thus, the x-drive signal S applied to the x-transmitter is Smax. sin(120°) ( 0.87 Smax), and the y-drive signal applied to the y-transmitter is Sy = Smax. cos(120°) ( -0.5 Smax). The minus here indicates that the signal applied to the y-transmitter should be inverted with respect to the signal applied to the x-transmitter (as well as being only 57.7% of the strength).

The combined resultant of the x-transmitter broadcasting with a dipole moment signal S ( Smax. sin(120°)) and the y-transmitter broadcasting with a dipole moment signal S, ( Smax. cos(120°)) may again be considered as being equivalent to a single "virtual" I-lED transmitter V. The magnitude of the virtual dipole S in Figure 4A is again S,,,. (i.e. S,2 = S2 + S2). The orientation Pv of the dipole is 120°. That is to say, the resultant virtual dipole V is arranged broadside to the receiver 25 at an azimuth angle 300.

Thus simply by applying different drive signals to the respective x-and y-transmitters, both TM dominated data (inline virtual dipole transmitter, as in Figure 4A) and TE dominated data (broadside virtual dipole transmitter, as in Figure 4B) may be obtained for an arbitrary receiver location without need to move / rotate the source 22.

It will be appreciated that by scaling-down both the x-and y-drive signals in the manner shown in Figures 4A and 4B, a constant virtual dipole magnitude (corresponding to Smax) may be provided regardless of the virtual dipole's orientation.

This can provide for more simple interpretation of data obtained for virtual transmitters in different orientations (because the different virtual transmitters all have the same strength). However, for determining the orientation of the virtual dipole it is the relative strengths of the dipoles with respect to one another that is significant, and not necessarily their absolute strengths. Thus, in some embodiments, e.g. to maxim ise the magnitude of the virtual dipole transmitter for each orientation, one or other of the transmitters could be provided with the maximum drive signal Smax, with only the other transmitter dipole being provided with a drive signal scaled down according to the desired orientation of the virtual dipole q%. -i.e. scaled by a factor tan(p) or 1/tan(cp) depending on which of the two dipole transmitters is provided with the maximum drive signal Smax. The dipole transmitter that is closest to the desired virtual transmitter orientation should be the one that is provided with Sm*x. This is because otherwise a signal greater than Smax would be required for the other dipole transmitter -22 -in order to obtain the desired orientation of the resultant virtual dipole. For example, if the desired orientation Pv is within 45° of the y-transmitter, the y-transmitter should receive the maximum drive signal S, and the x-transmitter should receive a signal scaled by a factor tan(q) (i.e. S, = Smax and S, = Sma,c. tan (pu)). If, on the other hand, the desired orientation q is within 45° of the x-transmitter, the x-transmitter should receive the maximum drive signal Sm, and the y-transmitter should receive a signal scaled by a factor l/tan() (i.e. S = Smax and S, Sm / tan (q)). An advantage of this approach is that it maximises the amplitude of the virtual dipole for any given desired orientation and so can provide for an increased signal-to-noise ratio in measured data. For example, the virtual dipole in Figure 4A has a strength S Sm.

However, with S, = Sm and S. tan (q), the same virtual dipole orientation of 30° is provided with a greater strength of S = 1.15 Sm. (i.e. sqrt(S2 + Sm2. tan2(p)).

Thus in a practical survey having multiple receivers at multiple azimuthal orientations with respect to the source 22, e.g. such as shown in Figure 2, virtual inline (TM dominated) and virtual broadside (TE dominated) data can be sequentially obtained for all receivers by sequentially adjusting the relative strengths of the drive signals applied to the respective transmitters in dependence on the azimuth of each detector with respect to the HED transmitters comprising the source in turn.

Figure 5 schematically shows in plan view another electromagnetic survey according to an embodiment of the invention. Figure 5 is generally similar to and will be understood from Figure 2, with corresponding elements identified with the same reference numerals. However, Figure 5 shows a different survey configuration. Thus the area surveyed / monitored comprises a section of seafloor 6 above a subterranean hydrocarbon reservoir 57 having a boundary 59. In this case ten receivers 25 are arranged generally over the reservoir 57. The source 22 in this example is located to one side of the reservoir. The source 22 and receivers 25 may be the same as those shown in Figure 2. Although Figure 5 shows a different survey layout to that shown in Figure 2, the principles according to which the two transmitters T, T comprising the source 22 are operated to provide a virtual dipole transmitter of any desired orientation are generally the same. An arrow 60 in Figure 5 schematically indicates the direction with which the virtual transmitter dipole is aligned. In Figure 5 the virtual transmitter is shown at an initial orientation of around p = 300, e.g. similar to the situation shown in Figure 4A.

It is assumed that the aim of the survey shown in Figure 5 is to obtain inline data from all receivers for the source and receiver placements shown. (With known survey methods this would require a conventional HED transmitter to be sequentially rotated about its axis so as to be inline with each receiver in turn such that inline data could be collected sequentially from each of the receivers.) The assumption made here that the aim of the survey shown in Figure 5 is to obtain inline data from all receivers for the source and receiver placements shown is purely to provide a concrete example to ease explanation. Survey configurations in accordance with embodiments of the invention may more generally be used to provide data corresponding to multiple virtual source orientations (e.g. inline, broadside, intermediate) for multiple source-receiver pairings. Significantly, data for the same virtual source orientations can be provided for each receiver regardless of its location.

In the initial orientation shown in Figure 5, none of the receivers 25 are inline with the virtual dipole. To survey the reservoir 57, the pointing direction (i.e. the inline direction) of the virtual dipole transmitter may be scanned over the detector array, as schematically indicated by arrow 62, by varying the relative strengths of the drive signals applied to the respective x-and y-transmitters appropriately. For example, the survey may involve scanning the virtual horizontal electric dipole transmitter from the initial orientation indicated by arrow 60 to a final orientation indicated by dashed arrow 61. E.g. the x-and y-drive signal strengths (i.e. amplitudes of waveform) may be varied linearly from their initial values associated with a virtual dipole orientation along arrow 60 to final values associated with a virtual dipole orientation along dashed arrow 61. This provides a scan of the area with the resultant virtual horizontal electric dipole transmitter so that virtual inline data are provided for all the receivers 25 at some time during the scan. The duration of the scan will depend on the amount of data required for the different orientations of the virtual horizontal electric dipole transmitter. For example, the scan might take several hours or days to complete, so that significant amounts of data are obtained from each receiver when in an inline orientation (e.g. when within 30° or so of the pointing direction of the virtual dipole transmitter).

-24 -In an alternative embodiment of the invention, the direction of the virtual dipole could "jump" between the individual directions. E.g. the x-and y-transmitters may be driven so that the resultant virtual dipole transmitters is directly inline with one receiver for a period of time, and then, discretely switched to being directly inline with another by appropriately changing the relative signal strengths applied to the respective transmitters.

It will be appreciated that the approach described above with reference to Figure 5 is fundamentally different from the rotating field approach described in WO 2006/059122 [7). The rotating field in WO 2006/059122 is provided by applying io signals of equal amplitude, but differing phase. This provides for a field that is elliptically / circularly polarised and which is not equivalent to a horizontal electric dipole source and so does not produce a geometric splitting. However, in accordance with the survey shown in Figure 5, the field is provided by applying signals of equal (or anti-) phase, but with differing relative amplitudes. This results in a selectively orientable virtual horizontal electric dipole (as apposed to the rotating non-linearly polarised field of WO 2006/059 122), and so maintains the geometric splitting effect that allows TE and TM mode data to be separated.

If desired, broadside data (or indeed data from any other desired orientation), could also be obtained for each receiver by appropriate directing of the virtual dipole.

When surveying with a virtual dipole transmitter in accordance with embodiments of the invention, any conventional electromagnetic parameter of interest may be recorded at the receiver for analysis (or derived later from raw or pre-processed receiver data). For example, amplitude and/or phase of a radial component of the electric field seen at a receiver when the virtual dipole transmitteris inline with the receiver may be obtained as the parameter(s) of interest, or an azimuthal component when the virtual dipole is aligned broadside, in line with conventional techniques for processing raw CSEM receiver data. Vertical electric fields could also be used. Other parameters often used as metrics in CSEM surveying relate to the polarisation ellipse described by the electric fields along two orthogonal horizontal directions at the receiver [1] (or polarisation ellipsoid in the case three-dimensional fields are measured at the receivers), and these too could be used. For example, the amplitude and / or phase of a semi-major axis of the horizontal polarisation ellipse

S

may be used as the metric of interest. Magnetic field based metrics may also be used in accordance with known techniques.

Data obtained using a virtual dipole transmitter, e.g. as described above, may be analysed in accordance with any conventional analysis techniques for CSEM data, that is to say, the data may be processed and analysed as if it had been collected by a "real" dipole transmitter having the characteristics (e.g. orientation, strength and phase) of the virtual dipole transmitter provided by the combined signals from the x-and y-transmitters. For example, using forward modeling, geophysical inversion and/or imaging/migration techniques.

Similarly, data may be obtained during surveying for various orientations of the virtual transmitter, for various ranges of source-receiver separations, and for various durations of data collection, as determined in accordance with conventional CSEM survey planning techniques -i.e. assuming the virtual dipole transmitter to be a real transmitter that may readily be aligned at any desired orientation.

In the example surveys in accordance with embodiments of the invention described above, receiver data are collected as the x-and y-transmitters are driven simultaneously by their respective drive signals. However, it is not necessary for the transmissions from the two transmitters to be simultaneous. This is because the signal couplings between the respective transmitters and the receiver are independent of one another. Thus, in another example survey, the x-transmitter may be driven by an x- drive signal S, during a first broadcast period (which may be referred to as an x-broadcast period) and a first data set D (which may be referred to as an x-data set) obtained at the receiver during this period. Subsequently, the y-transmitter may be driven by a y-drive signal S,. during a second broadcast period tt (which may be referred to as a y-broadcast period) during which a second data set D (which may be referred to as a y-data Set) IS obtained at the receiver. The x-and y-drive signals S, S, may be determined as described above (e.g. in dependence on a desired orientation of a virtual dipole transmitter). The only difference is that the signals are now applied sequentially with the x-and y-drive signals applied with a common phase with respect to their respective broadcast periods. Data from corresponding time bins in the x-and y-data sets may subsequently be added together to provide a combined data set which corresponds with a data set that would have been obtained if the x-and y-transmitters a -26 -were driven simultaneously (e.g. as shown in Figures 4A and 4B). The combined data set may then be processed, e.g. as described above, as if it had been obtained with simultaneous broadcast from both transmitters.

In the above-described examples, the orientation of the virtual dipole is s determined by applying an appropriate relative scaling to the drive signals applied to the x-and y-transmitters. However, the inventors have recognised that if the receiver signals from each of the two transmitters are separately measured at the receiver, e.g., by operating the two HED transmitters comprising the source 22 at different times, and collecting separate data sets at those different times, an effect corresponding to that of a virtual dipole having an arbitrary orientation can be achieved by applying a relative scaling to the receiver signals, rather than to the transmitter drive signals.

Figures 6A and 6B schematically show a manner of operating the source 22 described above for surveying in accordance with an embodiment of the invention based on scaling receiver signals instead of drive signals. Figure 6A shows the situation during a first broadcast period in which only the y-transmitter is driven and Figure 6B shows the situation during a second broadcast period in which only the x-transmitter is driven. The source 22 and receiver 25 shown in Figures 6A and 6B are assumed in this example to be in the same positions as the source and receiver seen in Figure 4A, and as with Figure 4A, it is assumed that there is a desire to measure TM mode dominated data at the receiver 25 (i.e. data equivalent to that which would be seen with a single dipole transmitter arranged inline with the receiver).

The schematic representations of the source 22 in Figures 6A and 6B are similar to, and will be understood from, Figures 4A. That is to say, in addition to schematically showing the respective physical x-and y-transmitters T, Ii,, Figures 6A and 6B also schematically show the strengths of the drive signals S,, S, applied to the transmitters during their respective broadcast periods. Thus during the first broadcast period (which may be referred to as a y-broadcast period), the drive signal applied to the y-transmitter is S, = Sm and the drive signal applied to the x-transmitter is S, = 0 (i.e. only the y-transmitter is broadcasting during the y-broadcast period). However, during the second broadcast period (which may be referred to as an x-broadcast period), the drive signal applied to the y-transmitter is S, = 0 and the drive signal applied to the x-transmitter is S, = Smax. The y-drive signal applied during the y-broadcast period and the x-drive signal applied during the x-broadcast period have the same waveform and phase with respect to their respective broadcast periods (which may be any conventional CSEM waveform).

A data set D (which may be referred to as a y-data set) is obtained during the y-broadcast period and represents the signal coupling between the y-transmitter and the receiver. Scaling the amplitudes of the measured electromagnetic field components comprising the y-data set D by a factor f provides a scaled data set D = The scaled data set D is equivalent to the data set that would be seen if the y-drive signal were scaled by the same factor f during transmission.

Data set D (which may be referred to as the x-data set) is obtained during the x-broadcast period and represents the signal coupling between the x-transmitter and the receiver. Scaling the amplitudes of the measured electromagnetic field components comprising the x-data set D by a factor f similarly provides a scaled data set = fD which is equivalent to that which would be seen if the x-drive signal S were scaled by the factor f during transmission.

Thus if the x-and y-data sets are scaled by respective factors f and f, and then summed (e.g. by adding the scaled measured EM field components seen in corresponding time bins in the respective broadcast periods), a combined data set is provided. This combined data set is equivalent to the data set that would be obtained with simultaneous transmission by the x-and y-transmitters with respective drive signals fxSmax and fySmax. Thus in accordance with the principles described above with reference to Figures 4A and 4B, the combined data set is also equivalent to the data set that would be seen for a single dipole transmitter having an orientation p = tan'(f/f).

The scaling factors f and f to apply to the x-and y-data sets may thus be determined in dependence on the desired orientation of a virtual dipole transmitter in the same way as the scaling factors for the x-and y-drive signals are determined in the embodiment described above in connection with Figures 4A and 4B.

For the receiver 25 shown in Figures 6A and 6B (i.e. at an azimuth of around (PR = 300), and assuming in this example there is a desire to provide a data set corresponding to that which would be seen for a transmitter dipole arranged inline with the receiver (i.e. p = PR = 30°), the y-data set should be scaled by a factor f = 0.87 (i.e. by sin((P)) and the x-data set should be scaled by a factor f, = 0.5 (i.e. by cos(qv)) and the resulting data sets combined. (This assumes the x-and y-drive signals have the same strengths during their respective broadcast periods. If they do not, this can also be accounted for in the scaling.) It will be noted the scale factors for the x-and y-data sets are the same scale factors as for the x-and y-drive signals applied to provide the same virtual dipole configuration in an embodiment such as shown in Figure 4A. Thus the general principles described above regarding how the x-and y-drive signals could be scaled relative to one another to provide the desired virtual dipole transmitter orientation also apply to how the x-and y-data sets could be scaled to provide data equivalent to that obtained from a virtual dipole transmitter having the io desired orientation. The only difference is that for embodiments such as shown in Figure 4A the scaling is applied at the transmission stage, whereas for embodiments such as shown in Figure 4B the scaling is applied at the detection stage.

In the general case for receiver-based scaling, for a source comprising two mutually orthogonal RED transmitters, a data set corresponding to that which would be seen for a virtual dipole having a desired orientation q may be provided by applying scaling factors to the respective x-and y-data sets so that the data sets are scaled relative to one another such that f/f = tan(). For example with f = sin(py) and f = cos(cpv). It is noted that again while in practice it may be simpler for the two RED transmitters comprising the source 22 to be orthogonal, they need not be. This is because any orientation of virtual dipole can be provided with appropriate receiver signal scaling for any two dipole transmitters that are not exactly parallel.

An advantage of the receiver-based scaling approach shown in Figures 6A and 6B over the source-based scaling approach shown in Figures 4A and 4B is that once the x-and y-data sets have been obtained, they may be scaled for any desired virtual dipole source orientation without requiring any further source transmissions. For example, assume now for the same source-receiver configuration shown in Figures 6A and 6B there is a desire to derive a combined data set which corresponds to data that would be seen for a transmitter dipole aligned broadside with respect to the receiver (i.e. for Pv = (pa + 900 = 1200). This might often be desired so that inline (TM dominated) and broadside (TE dominated) data for the same source-receiver configuration may be collected / compared I analysed. In this case the y-data set may 29 -be scaled by a factor f 0.87 (i.e. sin(l20°)), and the x-data set may be scaled by a factor f = -0.5 (i.e. cos(120°)), and the two scaled data sets combined. This combined data set is thus equivalent to that obtained with the source-based scaling represented in Figure 4B.

This means that although for a given desired virtual dipole orientation it takes twice as long to obtain data suitable for receiver-based scaling than for source-based scaling, the data suitable for receiver-based scaling can be differently combined to provide any number of corresponding virtual dipole transmitter orientations, whereas a retransmission is required to provide a different virtual dipole orientation using source-based scaling.

A further advantage of a receiver-based scaling approach is that the same strength drive signals may be applied to both the x-and y-transmitters. This means the signal-to-noise ratio is equal for transmissions from both transmitters. What is more, the two transmitters may each broadcast the maximum signal possible for the apparatus at hand, thus maximising the signal-to-noise ratio.

Another significant advantage of a receiver-based scaling approach is that the same drive signals are applied regardless of the azimuth of the receiver. This means multiple receivers at different azimuths can collect their own x-and y-data sets at the same time. The x-and y-data sets for each individual receiver may then be scaled and combined as appropriate to provide data corresponding to any desired virtual dipole orientation for that receiver.

Figure 7 schematically shows in plan view an electromagnetic survey layout that is similar to and will be understood from that shown in Figure 5. However, for the survey represented in Figure 5, a source-based scaling approach is used to scan or sequentially orientate a virtual horizontal electric dipole transmitter across the array of ten receivers 25. This allows virtual inline data to be obtained sequentially for each receiver. However, for the survey represented in Figure 7, a receiver-based scaling approach is used at each receiver to derive virtual inline data for all receivers simultaneously.

By way of example, it is assumed the aim of the survey shown in Figure 7 is to provide exactly inline data for each receiver for a twelve-hour period. The y-transmitter is first driven with the desired signal waveform (at maximum strength) for

S

-30 -a y-broadcast period of twelve hours, while the x-transmitter is not broadcasting. A resulting y-data set is obtained at each receiver during this time. Then, the x-transmitter is driven with the desired signal waveform (again at maximum strength) for an x-broadcast period of twelve hours with the y-transmitter not broadcasting. A resulting x-data set is obtained at each receiver during this time. (Because the x-and y-transmitters are driven at different times, it is noted the x-and y-data sets could be obtained using a single dipole transmitter that is rotated between x-and y-aligned directions instead of separate x-and y-aligned transmitters.) Virtual inline data sets may be derived independently for each receiver by io appropriately scaling and combining each receiver's x-and y-data sets in a manner depending on each receiver's azimuth. This may be done in accordance with the principles set out above in connection with Figures 6A and 6B. Thus the desired twelve hours of virtual inline data for all receivers is obtained in only twenty four hours of transmission (twelve hours per HED transmitter). This is because all receivers can record the data required simultaneously. This survey time compares favourably with the 120 hours plus a conventional CSEM survey would take (i.e. 12 hours for each of the ten receivers plus time to re-orient the transmitter). The scanning survey shown in Figure 5 would also require much longer as separate source transmissions (i.e. with different relative x-and y-transmitter scalings) are required for each receiver.

It will be appreciated various timings of transmission could equally be used for the survey shown in Figure 7. For example, the x-broadcast period could be undertaken first. In other examples shorter periods of x and y-transmission could be interspersed with appropriate data concatenation or stacking. For example, the y-transmitter could be driven for three hours with the x-transmitter switched off, and then the x-transmitter could be driven for three hours with the y-transmitter switched off, and this sequence may then be repeated three more times to provide the desired integration time. This interspersed approach may by preferred if there are concerns about systematic drifts in the apparatus stability, for example.

A significant factor in being able to apply a receiver-based scaling approach is an ability to independently measure receiver signals associated with each transmitter.

This is so that the two data sets can be independently scaled in dependence on the desired virtual dipole transmitter orientation before being combined. In the above -31 - example the x-and y-data sets are separated by transmitting from the x-and y-transmitters at different times. Other techniques could be used.

For example, the x-and y-transmitters could in principle be driven simultaneously with slightly different fundamental frequencies. The x-and y-data sets s could then be extracted from the raw data based on the differing frequency components in the recorded data.

Another technique for separating simultaneously broadcast signals from the x-and y-transmitters is based on interpolation between different frequencies of signal.

For example, the x-drive signal might comprise a square wave at a fundamental io frequency v (thus containing components at v, 3v, 5v, lv, etc.). The y-drive signal might comprise a square wave at a fundamental frequency 2v (thus containing components at 2v, 6v, IOv, 14v, etc.). The receiver signals associated with each transmitter could then be separated on the basis of their differing frequency contents.

Then an interpolation could be made between the v and 3v frequency components in IS the x-data set to provide a "virtual" 2v frequency component for the x-data set. This "virtual" 2v component could then be combined with the real 2v component in the y-data set (after appropriate scaling) to provide for a virtual dipole transmitter of desired orientation at a frequency 2v. The 5v and 7v components in the x-data set could similarly be interpolated to provide a virtual 6v component for use with the real 6f component in the y-data set, and so on. In practice the time-separation approach is also likely to be preferred to this approach because of the relative complexity of this approach and likely greater potential for introducing errors.

Data obtained using the crossed dipole source 22 to provide for a virtual dipole transmitter (for both source-and receiver-based scaling approaches) may not necessarily be exactly equivalent to data from a real dipole having the desired orientation. This is because the cross arrangement of the dipoles means that mathematically the 1-LED vector associated with the x-transmitter and the MED vector associated with the y-dipole transmitter do not strictly add to a virtual dipole transmitter centred on the middle of source 22 because the two vectors do not strictly add "end-to-end". However, the inventors have modelled the extent of the error and found it to be relatively small.

The error is zero when the desired virtual dipole orientation aligns with one of the x-or y-transmitters of the source (because signals associated with the other transmitter are scaled to zero to there is no issue of "end-to-end" addition). The error is greatest when the virtual dipole has a desired orientation furthest from one of the x-or y-transmitters (e.g. for Pv = 45°, 135°, 225° or 315°). The inventors have compared the amplitude and phase of data obtained with a modelled virtual inline dipole transmitter and a modelled real inline dipole transmitter. For typical survey conditions, e.g. for a source frequency of 0.25 Hz and source transmitter lengths of 30Gm, the maximum error at a range of 2km is around 2.5% in amplitude and less than 1° in phase for the semi-major axis of a horizontal polarisation ellipse. The corresponding errors for the vertical electric field component are less than 1% in amplitude, and again less than 1° in phase. At greater ranges the errors are smaller still. For example, the maximum error at a range of 8 km is under 0.2% in amplitude and less than 0.05° in phase for the semi-major axis of a horizontal polarisation ellipse, and less than 0.2% in amplitude and around 0.1° in phase for the vertical components. Broadly similar, but slightly smaller, errors are seen when comparing the amplitude and phase of data obtained with a modelled virtual broadside dipole transmitter and a modelled real broadside dipole transmitter.

The errors are small in practice because the I-lED vector mispositioning errors associated with a crossed dipole source (i.e. the divergence from true end-to-end HED vector addition) are on the order of half the length of the transmitters comprising the source, which is small compared to the typical source-receiver ranges used for CSEM surveying (data are not generally collected for ranges approaches the transmitter length as the receivers typically saturate at these offsets).

Thus any mispositioning error is likely to be acceptable in most cases. Even in cases where the error is of concern, it can be mitigated, for example, by reducing the lengths of the HED dipoles comprising the source (and increasing the current supplied to maintain the same overall dipole magnitudes). Furthermore, since the error is known and calculable, it can be accounted for in later analysis by appropriate modelling of the two transmitters comprising the source. However, accounting for the error in this way is likely to lead to large increases in data processing time using conventional CSEM processing algorithms (because the dipoles cannot be considered as point dipoles).

The above-described examples have primarily focussed on a source comprising a pair of orthogonal dipoles crossing at their midpoints. This configuration provides a S compromise in which the maximum mispositioning error seen over the full range of virtual dipole orientations 00 to 3600 is minimised. However, in some surveying I monitoring / appraisal geometries, a particular pair of source dipoles might have a preferred orientation range within a particular sector. For example, a fixed installation might include two or three sources -if one is located generally towards a western edge of the survey area, the majority of receivers might be orientated at an angle of between 00 and 180°, for example (where 0° is aligned with North, and 90° is aligned with East). A T-shaped source arranged sideways (i.e. a source comprising an East-West aligned first I-LED transmitter and a North-South aligned second HED transmitter, wherein an Eastward end of the East-West aligned transmitter meets a midpoint of the IS North-South aligned transmitter) could be used, i.e. in an "end-to-midpoint" configuration. This could be used to reduce errors in the preferred orientation sector between 0° and 180° (in conjunction with an increased mispositioning error in the sector between 180° and 360°). Errors in the East-West axis could be eliminated by keeping the amplitude of the signal applied to the first!-[ED transmitter signal constant. An L-shaped source (e.g. a source comprising an East-West aligned first 1-LED transmitter and a North-South aligned second HED transmitter, wherein an eastern end of the East-West aligned transmitter meets a southern end of the North-South aligned transmitter) could similarly be used reduce the error in a particular quadrant, e.g., a North-East quadrant.

The above description has focussed on a fixed source (e.g. as might be appropriate for a reservoir monitoring implementation). However, the invention may equally be applied to a towed source survey.

Figure 8 schematically shows a surface vessel 14 undertaking a towed CSEM survey according to an embodiment of the invention. The surface vessel 14 floats on the surface 2 of a body of water, in this case seawater 4. A submersible vehicle 19 carrying a source 42 in the form of two crossed HED transmitters is attached to the surface vessel 14 by an umbilical cable 16. The umbilical cable provides an electrical, -34.-mechanical and optical connection between the submersible vehicle 19 and the surface vessel 14. The source 42 is shown in more detail in Figure 9 and comprises an orthogonal pair of I-LED transmitters I,, T which are broadly the same, and may be driven in broadly the same way, as the correspondingly labelled transmitters of the fixed source 22 described above. The towed source 42 also comprises bracing elements 43 and conventional drogues / fins 44 to help maintain the horizontal crossed dipole arrangement of the source 42 as it is towed through the water 4. The source 42 is mechanically connected to the submersible vehicle 19 via a tow line and mechanical coupling 45. One or more conventional CSEM receivers 25 are located on the seafloor 6 and operate generally in the same way as described above, e.g., in relation to Figure 2 as the source 42 is driven to broadcast while being towed over the survey area of interest.

Figure 10 is broadly similar to and will be understood from Figure 2, but shows a survey in which the source 42 is towed over the reservoir 56 from a point marked F to a point marked G. Thus at any point along the tow path FG, the drive signals applied to the transmitters T, T comprising the source 42 may be scaled (in a source- based scaling approach), or the data from the receivers 25 may be scaled (in a receiver-based scaling approach) in accordance with the above-described principles to provide for any desired virtual transmitter orientation. What is more, this can be done for each receiver for a range of different offsets as the source is towed over the receiver array.

Thus one potential advantage of a towed scheme is that increased coverage in terms of the number of different source-receiver ranges sampled for a given number of receivers may readily be provided.

The general principles of the surveying methods described above for the fixed crossed HED transmitter source 22 embodiments apply equally to the towed source 42 of Figure 8. However one additional consideration when using a towed source might apply when the x-and y-transmitters are not driven simultaneously (e.g. for receiver-based scaling). In these cases it may be preferable to minimise the shift in source position between the times at which the x-transmitter and the y-transmitter are broadcasting to reduce errors that might be associated with the resulting x-and y-data sets sampling different regions of the earth, and with different offsets. This may be achieved in some examples by using an interspersed broadcasting approach having relatively short alternating x-and y-broadcast periods. In this case the average source position during transmission from the x-transmitter differs from the average source position during transmission from the y-transmitter by only the distance moved by the source in the length of one of the alternating broadcast periods.

Figure 11 schematically shows one way of driving the respective x-and y-transmitters of the source 42. The upper trace schematically represents the waveform of the drive signal S, applied to the x-transmitter and the lower trace schematically represents the waveform of the drive signal S, applied to the y-transmitter. The respective transmitters are driven to broadcast a square wave at a frequency 0.25 Hz in io alternating 16-second broadcast periods (i.e. four cycles per broadcast period). The drive signals applied to each transmitter are the same apart from the 16-second shift associated with thel6 second alternating cycles. Thus at each receiver, data obtained from times t = Os to 1 6s, from 32s to 48s, from 64s and 80s (and so on) comprise an x-data set, and data obtained from times t = I 6s to 32s, from 48s to 64s, from 80s and 96s (and so on) comprise a y-data set. Depending on the integration period required for an analysis window (which may be determined in accordance with usual practice), a series of 16-second x-broadcast periods may be concatenated / stacked accordingly to provided an integrated x-data set having the desired integration time. A corresponding series of 16-second y-broadcast periods may be similarly concatenated / stacked to provided an integrated y-data set. These two integrated data sets may then be analysed as described above, e.g. using a receiver based scaling approach. The average position of the x-transmitter and the average position of the y-transmitter during their respective broadcast periods only differ by the distance travelled by the source in 16 seconds. Furthermore this slight shift in the effective mean x-and y-transmitter positions could in any case be used to reduce overall positioning errors since the shift could to some extent balance the above-described mispositioning error associated with the not strictly "end-to-end" vector addition. For example, this could be achieved by arranging the tow speed and the duration of the alternating broadcast period so that the average position of the y-transmitter is near to an end of the x-transmitter when at its own average position.

Figure 12 is similar to and will be understood from Figure 11 but schematically shows another way of driving the respective x-and y-transmitters of the source 42.

Figure 12 differs from Figure 11 in that it makes use of the alternating "dead times" during which the respective transmitters in Figure 11 are not broadcasting to broadcast at another frequency. in the example shown in Figure 12, the other frequency is 0.125 Hz such that there are 8 cycles in each alternating 16-second transmission. Although the x-and y-transmitters broadcast simultaneously, the use of different frequencies allows the respective responses to the x-and y-transmitters to be separated at the receiver. The alternating periods of 0.25 Hz transmission may thus be stacked / collated to provide integrated x-and y-data sets for a 0.25 Hz signal. The alternating periods of 0.125 Hz transmission may similarly be stacked / collated to provide integrated x-and y-data sets for a 0.125 Hz signal. This approach thus allows virtual dipole transmitters of arbitrary orientation to be provided for using receiver-based scaling at both 0.125 Hz and 0.25 Hz. The approach thus makes more efficient use of the available transmitter time for each transmitter, and also increase the bandwidth of the survey.

is This dual-frequency approach could equally be applied to fixed-source surveys, and also to surveys using single extended broadcast periods for each transmitter (as opposed to interspersed alternating broadcast periods). For example, in a fixed installation where 12 hours integration is required, the x-transmitter may be driven at 0.25 Hz for twelve hours with the y-transmitter simultaneously driven at 0.125 Hz.

Subsequently the y-transmitter may be driven at 0.25 Hz for twelve hours with the x-transmitter simultaneously driven at 0.125 Hz. The receiver responses to the simultaneously broadcast 0.125 Hz and 0.25 Hz signals may be separated based on frequency content such that x-and y-data sets of 12 hours integration are obtained at both 0.125 Hz and 0.25 Hz. Again this allows virtual dipole transmitters of arbitrary orientation to be provided for at both frequencies in accordance with the above described receiver-scaling techniques.

In the foregoing description, the invention is described in terms of frequency domain manipulations of the transmitted signals and the recorded data. It will be understood that the method can equally be extended to time domain processing methods for both the source-based scaling and the receiver-based scaling implementations. A variety of conventional time domain processing techniques are well known to those versed in the art of time domain signal processing and any such techniques could be used. For example, Wright et al 2001 [10] and 2002 [11] describe time domain data processing techniques in the context of CSEM that could be used in conjunction with embodiments of the invention.

For source-based scaling in the time domain, two source dipoles which are s coupled to the waveform generator in parallel can be considered as having the same waveform, but with a scaling factor a in between: s2(t) czs1(t), dcending on the desired orientation of the virtual dipole. The two sources can be normalised in a way that preserves the scaling factor, and thus the orientation of the resultant virtual dipole.

The methodology described above for the frequency domain case can then be applied in the time domain.

Receiver-based scaling for source broadcasts separated in time similarly requires source normalisation prior to scaling of the received signals. The scaling may then be performed in the same manner as the frequency domain case described above.

For the time-domain application of the receiver-based scaling, with two sources emitting simultaneously, the invention may incorporate methodology for retrieval of the signal for each individual source from the combined signals detected at the receiver.

Various different methods may be used to separate the mixed source signals.

One such technique is described by Xia et al [12]. The technique of Xia et al is described in the context of simultaneous recording and later separation of two or more seismic signals, but the same methodology could be readily applied to CSEM time domain signals. For example Pseudo-Random Binary Signals (PRBS) could be used to generate source signals, rather than a frequency sweep such as used in the seismic applications of the methodology.

A possible, but not the only, implementation is to impose orthogonality on the transmitted signal from each source, by which is meant that differing source signals are either chosen or treated so as to have zero cross-correlation between one another.

Where the two source signals are orthogonal, they can be separated at the receivers by performing cross-correlation as follows.

The received signal r(t) is described by: rQ) = s () * em (,) + s2 () * em2 () where s1(t) is the signal emitted by a first source, and s2(t) is the signal emitted by a second source, as a function of time t. Here, " * " is the convolution symbol, and below " 0 " is the cross-correlation symbol. em1(t) is the diffusive model transfer function due to the first source (orientated in a certain direction), em2(t) is the diffusive model transfer function due to the second source (orientated in another direction).

Cross-correlating the received signal r(t) with s1(t) gives: s1(t)�r(t) = [s1(t)0s1(t)]"em1(t) because s1(t)�s2(i)=O by design.

Similarly cross-correlating the received signal with s2(t) gives: s2(t)Or(t) = [s2(t)0s2(t)]*em1(t).

The cross-correlation thus separates the signals from thc first and second sources, which were previously mixed. Each source term is now the autocorrelation of the individual source terms s1(t) and s2(t).

Because s1(t) and s2(t) are known by design, further processing (source is signature shaping) can be applied to normalise the differing resulting source terms, as is common in standard applications of the CSEM method. Each separated signal can be freely scaled according to the virtual dipole orientation and recombined as described for the frequency domain method.

It will be understood that embodiments of the invention are equally applicable to surveying in seawater or freshwater, for example large lakes or estuaries, so that references to seafloor, seawater etc. should not be regarded as limiting and should be interpreted as covering lakebed, riverbed etc..

Figure 13 is a schematic view of an oil rig 140 producing hydrocarbon according to an embodiment of the invention. The oil rig is located in the vicinity of a region of interest which has been CSEM surveyed in accordance with either of the above-described receiver-based or source-based scaling techniques. Here an analysis of the data obtained during the survey in accordance with embodiments of the invention has identified a subterranean hydrocarbon reservoir 12 within the region of interest. The identified hydrocarbon reservoir has been penetrated by a hydrocarbon-producing well 142 carried by the oil rig 140. Hydrocarbon (e.g. oil) may be produced -39 -from the well 142 (i.e. extracted / recovered from the reservoir 12) using conventional techniques.

Figure 14 is a schematic perspective view of a barrel containing a volume of hydrocarbon 144 according to an embodiment of the invention. The hydrocarbon is produced using the well 142 shown in Figure 13.

Figure 15 is a schematic perspective view of a data storage medium 146 bearing data obtained during surveying according to an embodiment of the invention.

The data storage medium in this example is a conventional optical disc, e.g. a data compact disc or data DVD disc. Any other storage medium may equally be used. Thus data sets obtained during surveying according to embodiments of the invention may be stored on the data storage medium 146 for later analysis.

Thus various controlled source electromagnetic survey methods for surveying subterranean strata are described. The methods rely on an electromagnetic source comprising two orthogonal horizontal electric dipole transmitters. Broadcast signals from the two transmitters combine to provide a corresponding resultant virtual dipole transmitter. The effective orientation of the resultant dipole transmitter may be arbitrarily selected so that data for arbitrary transmitter orientations may be provided (e.g. inline or broadside as desired). This is achieved without needing to move a source around, or to provide separate sources for each desired orientation. In some examples, a desired resultant virtual transmitter orientation may be provided by applying appropriately scaled drive signals to the two transmitters. In other cases, a desired effective resultant virtual transmitter orientation may be provided by applying equal drive signals to the two transmitters, but scaling the receiver signals associated with each transmitter relative instead. I.e. in some cases the relative scaling may be applied at the transmission stage, and in other cases the relative scaling may be applied at the detection stage.

-40 -

REFERENCES

[1] GB 2 382 875 (University of Southampton) [2] GB 2 411 006 (OFIM Limited) [3] GB 2 423 370 (01-EM Limited) [4] Constable, S. C. & Weiss, C., Mapping Thin Resistors And Hydrocarbons With Marine EM Methods: Insights From I-d Modelling, Geophysics, volume 71, issue 2, pp. 43 -51, March 2006 [5] GB 2415 785 (OHM Limited) [6] WO 2003/104844 (The Regents of the University of California) [7) WO 2006/059122 (ElectroMagnetic GeoServices AS) [8] WO 2004/053528 (The Regents of the University of California) [9] GB 2 415 785 (OHM Limited) [10] Wright, D.A., Ziolkowski, A. & Hobbs, BA, 200[, Hydrocarbon detection with a mu/tichannel transient electromagnetic survey, 71' Annual International Meeting, SEG, Expanded Abstracts, 1435-1438.

[11] Wright, D.A., Ziolkowski, A. & Hobbs, B.A., 2002, Hydrocarbon detection and monitoring with a multichannel transient electromagnetic (MTEM) survey, The Leading Edge, 21, 852-864.

[12] Xia, J., Geier, N. A., Miller, R. D., & Tapie, C. R., Orthogonal vibroseis sweeps, Geophysical Prospecting, 2005, 53, pp. 677-688 -41 -

Claims (32)

1. CLAIMS1. A controlled source electromagnetic survey method for surveying an area thought or known to contain a subterranean resistive or conductive body, the method comprising: providing a first dipole transmitter aligned with a first direction; providing a second dipole transmitters aligned with a second direction; obtaining a first data set at a receiver for a first source signal broadcast by the first transmitter; obtaining a second data set at the receiver for a second source signal broadcast by io the second transmitter; scaling one of the data sets relative to the other by a relative scaling factor determined according to a desired orientation of an effective resultant source signal corresponding to a combination of the first and second source signals; and combining the relatively scaled first and second data sets to provided a combined dataset.
2. 2. A method according to claim 1, wherein the first and second directions are orthogonal.
3. 3. A method according to claim I or 2, wherein the first and second source signals are broadcast at different times.
4. 4. A method according to claim 3, wherein the first and second source signals are broadcast in a series of alternating broadcast periods.
5. 5. A method according to claim 3 or 4, wherein the first and second source signals primarily comprise components at a first frequency, and the method further comprises obtaining a third data set at the receiver for a third source signal broadcast by the first transmitter at times when the second transmitter is broadcasting the second signal, and obtaining a fourth data set at the receiver for a fourth source signal broadcast by the second transmitter at times when the first transmitter is broadcasting the first signal, andS-42 -wherein the third and fourth source signals primarily comprise components at a second frequency different from the first frequency.
6. 6. A method according to claim I or 2, wherein the first and second source signals are broadcast simultaneously.
7. 7. A method according to claim 6 wherein the first and second source signals comprise different frequency components.
8. 8. A method according to claim 7, wherein the first source signal comprises a first fundamental frequency component and the second source signal comprises a second fundamental frequency component, and wherein the second fundamental frequency component is at a frequency that is an even multiple of the frequency of the first fundamental frequency component.
9. 9. A method according to claim 7 or 8, further comprising interpolating between data at two or more frequency components in one of the data sets to provide interpolated data at a frequency corresponding to a frequency component in the other data set.
10. 10. A method according to any of claims Ito 9, wherein the relative scaling factor is approximately equal to a tangent of an angle between the desired orientation of the effective resultant source signal and one of the first and second directions.
11. II. A method according to any of claims I to 10, wherein the desired orientation of the effective resultant source signal is selected according to an angular position of the receiver with respect to one of the first and second directions.
12. 12. A method according to any of claims I to 11, wherein the desired orientation of the effective resultant source signal is an inline orientation with respect to the receiver.
13. 13. A method according to any of claims 1 to 11, wherein the desired orientation of the effective resultant source signal is a broadside orientation with respect to the receiver.
14. 14. A controlled source electromagnetic survey method for surveying an area thought or known to contain a subterranean resistive or conductive body, the method comprising: providing a first dipole transmitter aligned with a first direction; providing a second dipole transmitters aligned with a second direction; applying a first drive signal to the first transmitter so as to broadcast a first source signal; and applying a second drive signal to the second transmitter so as to broadcast a o second source signal, wherein the relative strengths of the first and second drive signals are selected according to a desired orientation of a resultant source signal corresponding to a combination of the first and second source signals.
15. 15. A method according to claim 14, wherein the first and second directions are orthogonal.
16. 16. A method according to claim 14 or 15, wherein the first and second source signals are broadcast simultaneously.
17. 17. A method according to claim 14 or 15, wherein the first and second source signals are broadcast at different times.
18. 18. A method according to any of claims 14 to 17, wherein a ratio of the strengths of the first and second source signals is approximately equal to a tangent of an angle between the desired orientation of the resultant source signal and one of the first and second directions.
19. 19. A method according to any of claims 14 to 18, further comprising providing a receiver for detection of source signals broadcast from the first and second transmitters, wherein the desired orientation of the resultant source signal is selected according to an angular position of the receiver with respect to one of the first and second directions. S. -44 -
20. 20. A method according to claim 19, wherein the desired orientation of the resultant source signal is an inline orientation with respect to the receiver.
21. 21. A method according to c'aim 19, wherein the desired orientation of the resultant source signal is a broadside orientation with respect to the receiver.
22. 22. A method according to any of claims 14 to 21, wherein the relative strengths of the first and second drive signals are changed during the survey so as to cause a change in orientation of the resultant source signal corresponding to the combination of the first and second source signals.
23. 23. A method of monitoring a hydrocarbon reservoir, comprising performing a survey of an area known to contain the hydrocarbon reservoir according to any preceding claim at a first time, performing a survey of the area known to contain the hydrocarbon reservoir according to any preceding claim at a second time, and comparing results from the survey performed at the first time with results from the survey performed at the second time to identify changes in the reservoir between the first and second times.
24. 24. A computer readable storage medium having electromagnetic field data obtained during surveying according to any of the preceding claims recorded thereon.
25. 25. A method for obtaining hydrocarbon from an area that contains a subterranean hydrocarbon reservoir, comprising: penetrating the subterranean hydrocarbon reservoir with a hydrocarbon-producing well; extracting hydrocarbon from the subterranean hydrocarbon reservoir using the hydrocarbon-producing well; surveying the reservoir using a method according to any preceding claim in order to monitor depletion of the reservoir; and continuing to extract hydrocarbon from the hydrocarbon-producing well.
26. 26. A volume of hydrocarbon extracted according to the method of claim 25. -45 -
27. 27. A method of processing results from a controlled source electromagnetic survey of an area thought or known to contain a subterranean resistive or conductive body, the method comprising: providing a first data set obtained at a receiver for a first source signal broadcast by a first dipole transmitter aligned with a first direction; providing a second data set obtained at the receiver for a second source signal broadcast by a second dipole transmitter aligned with a second direction; scaling one of the data sets relative to the other by a relative scaling factor io determined according to a desired orientation of an effective resultant source signal corresponding to a combination of the first and second source signals; and combining the relatively scaled first and second data sets to provided a combined data set.
28. 28. A source for controlled source electromagnetic surveying, comprising: a first dipole transmitter aligned with a first direction; a second dipole transmitter aligned with a second direction; and a source control unit operable to provide first and second drive signals having selectable relative strengths to the respective first and second transmitters.
29. 29. A source according to claim 28, wherein the source control unit is further operable to determine relative strengths for the first and second drive signals according to a desired orientation of a resultant source signal corresponding to a combination of first and second source signals broadcast by the first and second transmitters in response to their respective drive signals.
30. 30. A controlled source electromagnetic survey method substantially as hereinbefore described with reference to Figures 2 to 15 of the accompanying drawings.
31. 31. A source for controlled source electromagnetic surveying substantially as hereinbefore described with reference to Figures 2 to 15 of the accompanying drawings.S
32. 32. A method of processing results from a controlled source electromagnetic survey substantially as hereinbefore described with reference to Figures 2 to 5 of the accompanying drawings.