Near-Surface Electromagnetic Surveying

Transcription

1 Near-Surface Electromagnetic Surveying The E&P industry typically focuses on deep formations, but frequently the nearsurface layers also need to be evaluated. Land-based electromagnetic surveys provide insights into this often complex zone. The interpreted resistivities of these layers help map and define features for applications as diverse as seismic studies and aquifer delineation. Mohamed Dawoud Environment Agency Abu Dhabi Abu Dhabi, UAE Stephen Hallinan Milan, Italy Rolf Herrmann Abu Dhabi Frank van Kleef Dubai Petroleum Establishment Dubai, UAE Oilfield Review Spring 2009: 21, no. 1. Copyright 2009 Schlumberger. For help in preparation of this article, thanks to Marcus Ganz, Houston. 1. The principle describing the changing magnetic field is Faraday s law of induction. Including the sign of the induced current is Lenz s law. The converse principle involving a changing current or electric field is Ampère s law. These laws are included in Maxwell s equations. The near surface of the Earth is a complex place, the result of dynamic action by wind, water and other forces of nature. In those first tens of meters below the surface, the jumbled detritus of weathering is gradually buried. The near-surface layers, like those below, vary in resistivity according to their mineral and fluid compositions. This property allows their investigation using electromagnetic (EM) surveys. Often, surveys are performed using an artificial source of EM radiation, rather than the magnetotelluric (MT) radiation resulting from the interaction of the solar wind with the Earth s magnetosphere. On land, there are two general methods of controlled-source electromagnetic (CSEM) measurement for generating the signal and detecting the response. The grounded-source method requires burial of source and receiver electrodes in electrical contact with the earth. The inductive-source method uses a current loop on the surface to induce a variable magnetic field, and the same or another loop to detect the response signal. The grounded-source method is efficient and sensitive to horizontal resistive targets because the electric field has a vertical component. The grounded receivers measure the response electric field; response magnetic fields are also measured to provide control during modeling. On land, however, the surface conditions must be suitable for creating and maintaining electrical contact. This prerequisite excludes the practical application of this method in large, arid dunes sand grains are nonconducting. But in some areas, contact can be improved by drilling patterns of shallow holes for source and receiver electrodes and wetting the soil as the hole is refilled. Deeper investigation into the Earth requires stronger current sources, among other factors, and the high contact resistances on land mean this requires high-voltage systems to drive that current. The inductive-source method does not require electrical contact, since the current loop generates a magnetic field through a time-varying signal. This field generates a response electric field, but because the electric field is largely horizontal, the process is not as efficient for imaging horizontal layers of resistive hydrocarbons as the direct injection of current using the grounded-source method. Again, both the electric and magnetic response fields can be measured using the inductive-source technique. Coils for the current loops are square and for nearsurface investigation range from about 10 to 300 m [30 to 1,000 ft] on a side. Far larger loops have been used for deeper, but low-resolution, investigation. A companion article (see Electromagnetic Sounding for Hydrocarbons, page 4) describes the basic physics of the EM interaction with the Earth and discusses marine EM studies. It also covers MT in detail, because the objectives of those studies are similar for land and marine environments. This article focuses on investigations 20 Oilfield Review

2 using the inductive-loop method for nearsurface imaging, illustrated by two WesternGeco cases from the United Arab Emirates. One study mapped an aquifer in Abu Dhabi for a waterstorage project. The second determined nearsurface resistivity variations in Dubai sand dunes, providing valuable input for making static corrections in a seismic survey of the area. Stacking Time Sequences Maxwell s equations describe the basic physics behind the interplay of electric and magnetic fields in a time-varying current loop. A current loop generates a magnetic field. If the current changes, the induced field also changes. The opposite is also true: Changing the flux of a magnetic field within a conducting loop induces a changing current. 1 A simple way to generate such a current is to move a magnet toward or away from a wire loop. The movement changes the flux through the loop, inducing a current. This current induces a response magnetic field oriented to oppose the change in flux through the loop caused by the moving magnet. No actual magnet is required for this effect to take place. One coil with an imposed timevarying current sets up a time-varying magnetic Spring

3 > Basic induction loops. An alternating current passing through a set of coils (blue) induces a cyclic magnetic field. When this field passes through a second set of coils (red), it induces a cyclic current in that circuit. Thus, energy passes from one circuit to another without a direct electrical contact. This is the basis for a transformer, which is a device that converts from an input voltage to a different output voltage by having different numbers of loops in the two coils. Time on Time off Ramp time Measurement period Time on Time Ramp time Transmitter current and primary magnetic field Induced electromotive force in nearby conducting layers Response magnetic field induced by eddy currents Receiver-coil voltage from response magnetic field field in response. Current is induced in a second coil positioned close enough to experience the changing flux. This is the configuration of a transformer (left). Energy passes from one circuit to the other through the changing magnetic field. One method that uses inductive measurement for evaluating the near surface is a time-domain electromagnetic (TDEM) survey. A conducting loop of wire set out in a square on the surface of the Earth acts as the first coil, and the second loop forms in conducting formations of the Earth itself. The primary magnetic field from the transmitter loop generates horizontal currents, called eddy currents, immediately beneath the loop. These currents induce a response field that can be detected at a surface receiver loop, but this field also travels farther into the subsurface, generating progressively weaker eddy current loops with larger radii and smaller response fields (below). 2 Batteries and signal electronics Source current loop Secondary magnetic field Eddy currents induced by field change Eddy currents at later times > TDEM inductive method. A large square loop is placed on the surface at the sounding site (top right). Passing a current pulse through this loop generates the primary magnetic field. The field induces a secondary, eddy current loop in the ground, as described by Faraday s law (middle right). This secondary current induces a response magnetic field, which can be recorded by a receiver loop on the surface. In this case, the same loop is used as source and receiver. The primary field decays with depth into the ground, generating response fields from each subsequent depth (bottom right). The timing of the signals and response fields is also shown (left). Depth The transmitter and receiver loops can be the same wire coil when an appropriate time sequence of current steps is applied, or coaxial but separate coils in a typical setup. In all inductive-loop TDEM surveys, the time sequence begins by turning the current on to a constant DC value. Sufficient time elapses for transient responses in the subsurface to decay. Then, electronics shut the current off in a rapid, controlled ramp, inducing a known electromotive force in the immediate subsurface. The transient electromotive force generates eddy currents, producing a secondary magnetic field that decays with time. The secondary field is detected by the receiver coil. After enough time has elapsed, the sequence repeats with opposite polarity. Stacking many repeated responses improves the signal/noise ratio. As the eddy currents move progressively deeper into the subsurface, the response field contains resistivity information about deeper layers. Fundamentally, the resistivity variation with depth determines the rate of decay of the transient; higher conductivity results in slower decay. Inversion of the stacked data from a TDEM sounding reveals the distribution of nearsurface resistivity. The measured signals are very small, so land surveys must consider and avoid, if possible, any sources of noise. Electric trains, power lines, electric fences, buried utility cables, pipelines and water pumps distort the local measurement; large temperature variations and wind impact stability; and variations in soil moisture and permeability affect uniformity. 3 TDEM is not commonly used directly in oil and gas exploration, although it has utility in evaluating surface statics for seismic studies. The method is applied widely for exploration within the mining industry, employing both terrestrial and airborne sources. It is also a tool for environmental and water-resource management, as shown in the first of the following case studies from the Middle East. Sounding for Water Storage A recent land-based EM application helped to locate potential water-storage sites in the UAE. 4 The Environment Agency Abu Dhabi (EAD) is managing a study for the government to evaluate storage plans for 30 billion British imperial gallons [136 million m 3, 36 billion galus] of fresh water in the northeast area of the Emirate. 5 The country wants a freshwater reserve for emer gency periods and to meet peak 22 Oilfield Review

4 summer demands. The water for this aquifer storage and recovery (ASR) project will be transported via pipeline from a water-desalination plant in the Emirate of Fujairah. EAD retained Schlumberger to identify and test a potential ASR site. This involved defining the subsurface storage zone and surrounding formations, aquifer thickness and related hydraulic parameters. Schlumberger selected a preferred site and constructed three pilot wells, which were tested to determine the aquifer s potential. Geologic studies of the area found that deep-seated faults had been reactivated in the Late Tertiary Period by the northeastward displacement of the Arabian Peninsula, resulting in a series of tightly folded faults. The overlying layer of Quaternary Period sediments, consisting of eolian sands and alluvium, were primarily deposited along the reactivated faults and in the synclines between them, giving the sediment thickness a directional bias. The directionality can influence groundwater flow, creating a preference for flow parallel to the structure. Schlumberger evaluated this geologic structure in 2006 during the drilling, logging and testing periods. Logs from the pilot wells indicated a resistivity contrast between the targeted sand-and-gravel aquifer and the underlying clayrich layer. Because TDEM data are useful for aquifer characterization, the evaluation over the ASR site included a survey to define the lateral extent of the aquifer necessary to compute the potential water-storage volume. In a time-domain survey, the apparent resistivity of the underlying formations is determined from the time variation of the electric and magnetic response fields. For the ASR study, the same set of coils placed on the surface was used for both current and receiver loops. The survey covered a 6- by 7-km [3.7- by 4.4-mi] area. A 1D Occam inversion at each receiver yielded resistivity input for constructing a 3D model. 6 The maximum depth obtained by the inversion was about 250 m [820 ft]. Resistivity logs from the three wells were compared with the inversion results at adjacent sounding locations (right). The top of the underlying clay unit is clear in the eastern part of the survey area but less obvious in the western part. Sounding S049 Well SWS17 Sounding S071 Well SWS15 Sounding S013 Well SWS16 TDEM Resistivity Log Resistivity TDEM Resistivity Log Resistivity TDEM Resistivity Log Resistivity 1 ohm.m ohm.m ohm.m ohm.m ohm.m ohm.m 100 Depth, m Depth, m Bottom of aquifer Water table Water table Bottom of aquifer > Comparison of TDEM soundings with resistivity profiles from logged wells. The TDEM resistivity measurements from soundings correlate closely with the resistivity logs from adjacent wells. The soundings at S049 and S013 show reasonable correlation for the contact between the aquifer and the clay-rich unit below (violet), which is true for most other soundings in the eastern part of the investigated area. The contrast is not as clear at S071, where the more resistive lower unit does not provide a sufficient contrast for the TDEM measurement. This trend follows for most of the soundings in the western part of the survey. The top of the water table (blue dashed line) was determined from a map of the water depth by interpolating between wells in the area Water table Bottom of aquifer Nabighian MS: Quasi-Static Transient Response of a Conducting Half-Space An Approximate Representation, Geophysics 44, no. 10 (October 1979): Constable SC, Orange AS, Hoversten GM and Morrison HF: Marine Magnetotellurics for Petroleum Exploration, Part I: A Sea-Floor Equipment System, Geophysics 63, no. 3 (May June 1998): For more on water storage: Black B, Dawoud M, Herrmann R, Largeau D, Maliva R and Will B: Managing a Precious Resource, Oilfield Review 20, no. 2 (Summer 2008): A British imperial gallon is equivalent to 1 galuk. 6. An Occam inversion is a smooth inversion that does not predefine the number of layers. Spring

5 2 km 80 m 2 km > Resistivity discontinuity in the bottom of an aquifer. The discontinuity is a band with 15- to 20-ohm.m resistivity (yellow and green), which contrasts with the 1- to 10-ohm.m resistivity (blue and violet) elsewhere in the resistivity cube. The resistivity of the clay at the bottom of the aquifer (violet) to the east of the discontinuity is lower than that to the west. This discontinuity roughly aligns with a thrust fault (tan) that was identified at about 3,000 m [9,800 ft] by a seismic interpretation. Outcrop studies performed to the south of the survey area (not shown) support the surface expression of the fault being slightly west of the deeper seismic interpretation, and those observations are consistent with the location of the near-surface resistivity discontinuity. The syncline axis (blue) from the seismic interpretation, another thrust fault (purple) and some wells are also shown. (Seismic lines adapted from Woodward and Al-Jeelani, reference 7.) N The TDEM data clearly show a discontinuity in the resistivity distribution in the clay unit (left). There is also a difference in resistivity between the eastern and western compartments; the western part exhibits significantly greater resistivity at a given depth. The anomaly aligns with a seismically mapped thrust fault. 7 The seismic interpretation was based on a survey acquired in the early 1980s and reprocessed in 1992 to highlight the shallow structures. The shallow units above the anomaly are expected to show some structural complexity; they exhibit rapid variations in saturated thickness and possibly no saturated thickness in some areas. On the eastern side of the discontinuity, the horizontal extent of saturated thickness is suitable for an ASR unit. The western side of the discontinuity shows some potential, but has a larger risk: The interpretation in that area has a greater uncertainty because of the poor resistivity contrast between the saturated sands and the underlying clay. The discontinuity in the clay formation should not be seen as a complete hydraulic barrier in the shallower aquifer layer. Paleochannels or tear faults those striking perpendicular to the overthrust fault are expected to provide preferential flow paths from east to west across the line of the discontinuity. This TDEM study suggests that around 4 billion imperial gallons [18 million m 3, 4.8 billion galus] of water can be stored at this site, giving it the capacity for daily production of more than 20 million imperial gallons [91,000 m 3, 24 million galus] for 200 continuous days. Mapping the Dunes Within the same regional setting as the waterstorage site, a 2D seismic survey was conducted for Dubai Petroleum Establishment (DPE). The same clay layer forming the base of the storagesite aquifer is present in this area; it provides a marker for the base of the weathered surface layer. The depth of the clay varies across the survey area, and lines of dunes add local variation 7. Woodward DG and Al-Jeelani AH: Application of Reprocessed Seismic Sections from Petroleum Exploration Surveys for Groundwater Studies, Eastern Abu Dhabi, UAE, paper SPE 25538, presented at the SPE Middle East Oil Show, Bahrain, April 3 6, An uphole is a shallow well used to determine nearsurface velocities for a seismic survey. 9. Colombo D, Cogan M, Hallinan S, Mantovani M, Vergilio M and Soyer W: Near-Surface P-Velocity Modelling by Integrated Seismic, EM, and Gravity Data: Examples from the Middle East, First Break 26 (October 2008): > Desert terrain at the ASR site. The blue cable is part of a TDEM sounding loop. The building houses a submersible pump and water tank for the nearby wells (blue caps) of the ASR project. 24 Oilfield Review

6 Elevation, m North UH Distance, km Resistivity, ohm.m > Soundings along a seismic line. Interpolation between sounding points yields a detailed 2D model of resistivity along a seismic line. The sharply layered model at each sounding point (filled black square) is shown as a narrow column (top). Uphole sites (UH-09, -06, -08) contain domains of constant velocity in the weathered layer (yellow stippling) of about 1,400 m/s [4,600 ft/s] and in the underlying clay and limestone (gray stippling) of greater than 2,000 m/s [6,560 ft/s]. The variation in properties both within the dunes and in the lower layer is evident along the entire seismic line (bottom). The higher resistivity of the bottom layer (south end of the seismic line) indicates a relatively clay-poor region. to the depth of the weathered layer. Sand dunes generally exhibit low seismic velocity, and defining the velocity variation and thickness of the surface layer is crucial for deriving a long-wavelength static correction for the seismic data. The seismic crew drilled several upholes to log the velocity of the surface and underlying clay layer. 8 The upholes typically were at seismicline intersections and, for practical reasons, away from the higher dune crests. However, this pattern often does not sample the near-surface variations found in sand dune areas, so finer sampling was desired. DPE elected to use a TDEM resistivity survey to map the area because it would be more cost-effective than drilling more upholes and would avoid additional drilling on the environmentally sensitive dunes. UH-06 UH-08 South The survey comprised 505 sounding sites using square loops of 50 m [164 ft] on a side, except for a few sites that used 75-m [246-ft] square loops for deeper penetration. 9 Spacing between sounding points was generally about 1,000 m [3,280 ft]; GPS was used to position the sites. The effective time for decay ranged from 0.01 to 10 ms; the pulse repetition rate was 6.3 Hz. Given the subhorizontal nature of the zone of investigation and its shallow depth compared with the TDEM station spacing, 1D resistivity inversion modeling was selected for the analysis. Two 1D resistivity inversion methods were applied. The first one incorporated about 15 layers extending to a depth of about 200 m [650 ft]. Layer thickness increased logarithmically with depth. Resistivity was a free parameter, and this inversion yielded a detailed, smooth variation of resistivity The detailed fit provided a starting point for the second inversion. Termed a layered fit, it used the minimum number of layers required to fit the data to less than 5% root-mean-square misfit. This was typically two to five layers. Analysts selected the starting definition of these layers from the detailed fit. The layered model generated stronger resistivity contrasts than the detailed one. Interpreters created a 2D model with grid blocks 200 m wide and 5 m [16 ft] deep along a seismic line. They used the sounding sites along the seismic line to evaluate the surface layering. Model resistivity values were obtained by interpolating between the 1D smooth inversions at those sounding sites (left). The result is a detailed description of the location of the bounding clay layer at the bottom of the lowvelocity zone. The resistivity data were not calibrated to seismic velocities. The seismic processing team used these maps during the estimation of the surface static corrections. The velocities for the surface zone were interpolated from velocity measurements taken at the upholes. The TDEM approach gave the seismic interpreters a geologically consistent way to remove the effects of the laterally varying sand velocities. The resistivity analysis also highlighted variations within and below the lowvelocity weathered layer. Sounding Deeper Both case studies used TDEM methods to examine near-surface features. However, because using inductive-source techniques is inefficient for defining deep targets, the method is not the exploration tool of choice for examining deeper structures. The industry is improving techniques for using the alternative, grounded-source method to inject current into the earth. The source for the grounded method must be able to inject a large current at a voltage that is sufficient to overcome contact resistance in areas where the soil is dry. This combination has been difficult to achieve. The method using direct injection of current is more sensitive to resistive targets, making it a more likely method than the inductive-loop option to provide a direct hydrocarbon indication. MAA Spring