A novel use of marine controlledsource electromagnetic sounding tech-niques (CSEM), called seabed logging,may cut exploration costs in deepseaareas. The method has been tested offWest Africa in 2000 and in the Norwe-gian Sea in 2001 with promising re-sults.

Statoil has recently released informa-tion from ongoing research on electro-magnetic (EM) methods for deepseaprospect explo-ration whichshows that incertain areas itis possible todecide whethera deepsea struc-ture, mappedby seismic, has high resistivity or not.1 2

That is, the method may aid to deter-mine the probability for the structure tocontain hydrocarbons.

The enthusiasm in Statoil has trig-gered the establishment of a new com-pany dedicated to develop the methodfurther and to service the demand formarine EM measurements in the future.Statoil Innovation and one of Statoil’smajor research partners within geophys-ical EM research in Norway, the Norwe-gian Geotechnical Institute, own thenew company, Electromagnetic Geoser-vices (EMGS).

Here we present the method, firstwith an introduction covering the geo-physical concepts in general and quali-tative terms, and later a more theoreticalpresentation including calculations andnumerical computation for simplemodels.

IntroductionThe seabed logging (SBL) method is

a remote resistivity sensing methodwhich exploits the fact that hydrocar-bons are electric insulators and conse-quently, the hydrocarbon filled reser-voirs normally are more resistive thansurrounding water-filled sediments(Fig. 1).

The actual resistivity profile of thesubsurface is of course usually morecomplex than illustrated in the figure,and the resisitivity of the reservoirs willnormally vary depending on the porefluid composition (for example, the de-gree of hydrocarbon saturation).

Formulas derived from the classicalMaxwell’s equation show that the propa-gation speed and attenuation of a lowfrequency EM signal in a conductive en-vironment are determined by the resis-tivity of the medium and the frequencyof the EM signal. For a given frequency,a high resistive hydrocarbon filled reser-voir will represent a major positive elec-tric impedance contrast, giving rise toboth reflections and refractions.

Like the two vector modes of seismicS waves (SH and SV) which respond dif-ferently in layered sediments, EM waveshave the transverse electric (TE) andtransverse magnetic (TM) modes.The TEand the TM modes respond differently toa resistive layer (i.e., a hydrocarbon-filledreservoir) and this is utilized in the pro-cessing and interpretation of SBL data.

Despite the high impedance contrastimposed by a resistive layer, the TMmode penetrates into the layer through anarrow aperture of angle of incidents,where the reflection coefficient is small.The TM mode is trapped inside the layer

and propagates with higher speed andrelatively low attenuation throughoutthe resistive layer, both as “normalmodes” and “guided modes” (see latersection).The energy leaks out on theway, as in seismic refraction.

The method has recently been testedwith promising results in two hydrocar-bon provinces, off West Africa in 20003

and in the Norwegian Sea in 2001, by

‘Seabed logging’: A possible direct hydrocarbonindicator for deepsea prospects using EM energy

E X P L O R A T I O N & D E V E L O P M E N T

F.N. KongH.WesterdahlNorwegian Geotechnical InstituteOslo

S. EllingsrudT. EidesmoS. JohansenElectromagnetic Geoservices ASTrondheim, Norway

Exploration

ReportS P E C I A L

New Views of the Subsurface

Reprinted with minor edits, from the May 13, 2002 edition of OIL & GAS JOURNALCopyright 2002 by PennWell Corporation

using available scientific oceanographicmarine CSEM equipment made for otherapplications: a deep-towed, high currentelectrical dipole EM source, typically300A, 100 m long, supplied bySouthampton Oceanography Centre(SOC)4 and multiple ocean bottom re-ceiver stations supplied by Scripps Insti-tute of Oceanography.5 Fig. 2 illustratesthe principles. Typical source-receiveroffsets are one to five times the expectedreservoir depths.

One of the reasons for the low atten-uation of the TM mode inside reservoirsis that the high resistivity (low conduc-tivity) hinders the current to flow verti-cally between the top and base of thereservoir. At the reservoir boundaries, orat “holes” in the reservoir, the currentagain will flow as normal, and togetherwith internal reflection from the edges,a new anomaly will appear on the sur-face as “edge effects.”These effects maybe used to determine reservoir bound-aries.

The main factor that limits the use ofEM techniques as a high-resolutionstructural mapping tool of marine sedi-ments is the high attenuation of thehigher frequencies as a function of dis-tance.The key issue of the research hastherefore been to see if valuable infor-mation still can be extracted from EMdata even if the signal is low in frequen-cy and magnitude.

The answer is yes in many cases, asthe EM response on and off a deepseastructure will be significant and withinthe detection limits of available technol-ogy. It is the difference in response onand off a seismically mapped structurewhich would be used as a possible di-

rect hydrocarbon indicator, but otheroptions such as the difference betweenthe TE and TM mode are also important.

The magnitude of the EM responsedepends on the resistivity of the reser-voir, the burial depth, the reservoir sizeand continuity, and the electrical proper-ties of the overburden.The response islargest for shallow reservoirs with a highresistivity contrast to the overburden.Reservoir thickness also influences thedata but is considered less important(see later section).

With the current scientific data ac-quisition equipment, which is not tai-lored for the SBL application, one nor-mally would guess that the maximumexploration depth is approximately2,000 m below the sea surface. A simplemodeling exercise is normally carriedout prior to each specific survey to de-termine the expected maximum explo-ration depth. Deepsea is required in or-der to get low-noise recording condi-tions, as the seawater acts as a shieldfrom external atmospheric (magnetotel-luric) noise, and in order to avoid theinterference from air-waves (waves gen-erated by the source that refract andpropagate in the air).

Evaluation, examplesIn this section we will present some

basic theory and give calculation andmodeling examples to give the qualifiedreader an introduction to the most im-portant parameters and topics of the SBLmethod.

As a basis for the later discussions,consider the simple model shown in Fig.3:

• Overburden is 1,000 m thick with

resistivity R = 1 �m and relative permi-tivity � = 20

• A hydrocarbon reservoir, 50 mthick with resistivity R = 50 �m andrelative permitivity �=6

We will present the analysis step bystep, starting with an analytical homoge-nous model and ending with showingresults from a numeric 3D modeling ex-ercise.

Plane wave in a homogeneous model. The be-havior of a sinusoidal plane EM wavewith angular frequency � in a homoge-nous model with electric resistivity R,dielectric permitivity �, and magneticpermeability µ0 , can be described bythe complex wave number k:

k = �[ µ0 ( � + j/(R�) )]1/2 = � + j� (1)

where � is the phase constant (deter-mines the velocity) and � is the attenua-tion constant.

Figs. 4 and 5, respectively, show thephase velocity and the propagation at-tenuation in a model with resistivityR=1�m (overburden, Fig. 3) and withresistivity R = 50 �m (reservoir, Fig. 3)for the frequency band 0.1-5 Hz. Notethe substantial difference in velocity andattenuation for the two different models.Note also the dispersion, i.e. velocityvaries with frequency.

E-field produced by a horizontal dipole trans-mitter on the seabed. A plane wave and a ho-mogeneous medium assumption is ofcourse not valid for the SBL measure-ment situation in practice.

To approach the reality a bit more,consider a homogenous seabed model,like the one in Fig. 3 but without a hy-drocarbon-filled reservoir. Consider an

E X P L O R A T I O N & D E V E L O P M E N TSCHEMATIC ELECTRIC RESISTIVITY PROFILESCHEMATIC ELECTRIC RESISTIVITY PROFILE Fig. 1

Section contains water-saturated (W) and hydrocarbon-saturated (HC) sediments.Source: Statoil

Shale

HC-sand

W-sand

Shale

MEASUREMENT PRINCIPLEMEASUREMENT PRINCIPLE Fig. 2

Vessel tows a high-power electromagnetic source while recording the direct, reflected, and refracted signals on the seabed.Source: Statoil

horizontal electric dipole transmitterplaced on the sea floor.The complete ra-

dial elec-tric fieldE� (thefield alongthe anten-

na direction) in the seawater near thesea floor at a radial distance � (or z) canbe written:6

E� = I Lt ejkr ( 1/ �3 - jk/�2 - k2 /� ) /

(2��sea ) (2)

where I is the transmitter current, Lt isthe transmitter antenna effective length,k is the complex wave number of theseabed, � is the distance to the transmit-ter antenna, and �sea is the electric con-ductivity of the sea water.

The signal magnitude (Vr ) receivedby a horizontal dipole receiver antennaat a radial distance � (from the transmit-ter) can then be written

Vr = I Lr Lt ejkr ( 1/ �3 - jk/�2 - k2 /� ) /

(2��sea ) (3)

where Lr is the receiver antenna effectivelength.

Both the magnitude and the phase ofthe received signal, as functions of the

distance �, are measurable quantities.They are mainly dependent on theseabed wave number k. Fig. 6 shows thecurve of the magnitude of the receivedsignal, as a function of distance, normal-ized by the signal received at 100 m forthe frequency 1 Hz.

EM energy guided by a continuous high resis-tivity hydrocarbon reservoir. Refer to Fig. 3,where the horizontal direction (thelength direction) is defined as z or �, thevertical direction (the depth direction)as y, and the x direction (the width di-rection) is normal to y-z plane.

When there exists a hydrocarbonreservoir, the received signal containsthree parts: the received signal due tothe direct wave, which is the same as thesignal received at the homogeneousseabed case, the reflected wave, and therefracted wave. Since the reflected wavetravels a longer distance in seabed—ahigh attenuation material, the reflectedsignal is always weaker than the directwave. Hence it has less importance toour problem concerned. In the later dis-cussion, we mainly consider the refract-ed wave.

At first we consider the problem ofwave propagation inside the hydrocar-bon layer. Since the conductivity of theseabed, which surrounds the hydrocar-bon layer, is high, one can approximate-

ly consider the hydrocarbon layer as arectangular wave-guide. In practice, ofcourse, a hydrocarbon reservoir is acomplex structure, not a box. So this isonly for illustration purposes.

According to the theory of waveguide,7 the wave number in z directionkz can be written:

kz = (k2 – (m�/a)2 – (n�/b)2 )1/2

where m and n are integer numbers todefine the different propagation modesand a and b are the wave guide widthand height, respectively (Fig. 7).

For the main mode in consideration,m is 1 and n is 0, and

kz = (k2 – (�/a)2 )1/2

From the above equation one can de-rive the following:

kz is always smaller than k.Thus thephase velocity in the wave-guide of hy-drocarbon layer is always faster than thephase velocity of the wave traveling inan infinite hydrocarbon material with-out boundary.

When the width ‘a’ is much biggerthan the wavelength in hydrocarbon,then kz approaches k.

When k is smaller than �/a, then kzbecomes imaginary.That means the rec-tangular wave-guide works as a high-pass filter and has a cutoff frequency.7

However the wave is not completely cutoff. The wave-guide introduces a propa-gation attenuation when the frequencyis lower than the cutoff frequency.

kz is not related to b: the height ofthe hydrocarbon layer.That means thepropagation of the refracted wave insidethe hydrocarbon reservoir is not muchaffected by the height of the reservoir.This may also mean that the method canbe used to detect thin hydrocarbonreservoirs.

In Fig. 7, the EM field of the mainmode of a rectangular guide is shown.

From Fig. 4, one can see the strongcontrast between the velocities in theseabed and the hydrocarbon layer. Hencethe injected wave (Ray 1 in Fig. 3) forgenerating the refracted wave is almostnormal to the hydrocarbon layer, in or-der to generate wave propagating insidea thin hydrocarbon layer. Considering

WAVE PROPAGATION PATHS IN A HYDROCARBON RESERVOIR Fig. 3

Direct wave

Ray 1

1,000 mReflected

waveRefracted

wave

Y Z

Tx Rx

50 m

Sea0.2

ohm-m

Overburden(seabed)

1.0 ohm-m

Hydrocarbonreservoir

50 ohm-m

Sea floor

Seabed

New Views of the Subsurface

E X P L O R A T I O N & D E V E L O P M E N T

this fact, the received refracted signalVrefracted, can be written:

Vrefracted = A ei2kd e ikz� (4)

where A is a constant to characterize theantenna directivity, reflection coefficientetc., d is the depth of the hydrocarbonlayer, kz is the wave number in the waveguide formed by the hydrocarbon layer.The term ei2kd means that the refractedwave travels twice of the depth betweenthe sea floor and the hydrocarbon layer,and the term eikz� means that the refract-ed wave travels an additional distance �(the distance between the transmitterantenna and the receiver antenna) insidethe hydrocarbon wave guide.

The curve of the magnitude of the re-ceived refraction wave, normalized bythe refracted wave at 100 m, is alsoshown in Fig. 6 (line with symbol ‘+’).

From Fig. 6, one can see that the mag-nitude curve of the received direct waveis quite different from the refractedwaves.The magnitude of the received di-rect wave decays much more rapidlywith increasing the detection distance,than the refracted wave. Hence this phe-nomenon can be used as an “indicator”of the existence of the hydrocarbon layer.

We should note that the direct waveand the refracted wave curves shown inFig. 6 are all normalized with their ownvalue at � =100 m.

So far we haven’t touched the prob-lem about the magnitude of the receivedrefracted wave, comparing with the di-rect wave, although we have good rea-son to believe that the refracted wavewill be bigger than the direct wave,when the detection distance � is suffi-ciently large, since the refracted waveonly travels 2d distance in seabed, and

travels the rest distance in a resistive hy-drocarbon layer.The numerical simula-tions dis-cussed inthe nextsectionshow thatthe refracted wave becomes bigger thanthe direct wave when the detection dis-tance becomes about 2.5 d.

Example of 3D-FDTD EM modeling, studyingthe effect of reservoir width. A 3D finite differ-ence time domain (FDTD) software forEM-wave propagation has been devel-oped at the Norwegian Geotechnical In-stitute and used for calculating the fol-lowing layered model, which is isotrop-ic with horizontal layers (Fig. 8):

Layer 1: Seawater, thickness 540 m, re-sistivity 0.3 ohm-m.

Layer 2: Seabed (“overburden”), thick-

PHASE VELOCITY Fig. 4

104

5

5

4.5

4.5

4

4

3.5

3.5

3

3

2.5

2.5

2

2

1.5

1.5

1

1

.5

.50

0

Ph

ase

velo

city

, m/s

Frequency, Hz

x

Hydrocarbon layer

Seabed

PROPAGATION ATTENUATIONPROPAGATION ATTENUATION Fig. 5

40

5

35

4.5

30

4

25

3.5

20

32.52

15

1.5

10

1

5

.50

0

Att

enu

atio

n, d

B/k

m

Frequency, Hz

Hydrocarbon layer

Seabed

0

–50

–100

–150

RECEIVED SIGNAL MAGNITUDE VS. DISTANCE1RECEIVED SIGNAL MAGNITUDE VS. DISTANCE1 Fig. 6

1Normalized by the received signal at 100 m distance.

0 1,000 2,000 3,000 4,000 5,000

Mag

nit

ud

e, d

B

Distance, m

Direct wave

Refracted wave

Y

b

X

Z

aE field: H field:

HYDROCARBON LAYER AS A RECTANGULAR WAVE GUIDE Fig. 7

Frequency, Hz

New Views of the Subsurface

ness 1,000 m, resistivity 1.5 ohm-m.Layer 3: Reservoir, thickness 120 m,

resistivity 1.5 ohm-m when water filledand 50 ohm-m when filled with hydro-carbons.

Layer 4: As layer 2.

Thesource is a10 HzRicker sig-nal excit-

ing the Ez or E� component (E fieldalong the line).

The model dimension was 72 x 72 x36 grid cells. The grid was regular witha size of 60 m, i.e. the reservoir is onlyone grid cell thick.Time step was 4 mi-croseconds, and a total of 200,000 timesteps were run.

The receivers, or observation points,are located along a horizontal line, onegrid cell below the sea floor, 22 all to-gether.The receiver separation is 360 m(6 grid cells) and the first receiver is lo-cated at the source.The model contains a5 cell wide PML (perfect matching layer).

The reservoir length is 4.32 km andcovers the whole length of the model.Three different reservoir widths are test-ed: 3 km, 2 km, and 1 km.

Figs. 9 and 10 show the results, the Ezcomponent of the wave field at the men-tioned receiver locations.

The time histories are recorded from 0to about 500 milliseconds.The blacktraces are results from when the reservoir

“is filled” with hy-drocarbons (resis-tivity = 50 ohm-m) and red traceswhen it is filledwith water (1.5ohm-m).The tracesare normalized.

As we can seefrom the figures,the black tracesstart to have highermagnitude than thered traces when thedistance is biggerthan 2,400 m.Thisverifies that the re-fracted wave will bebigger than the di-rect wave, when thedetection distance �becomes large.

For the 3 kmwide reservoir, thefar-offset signalsare more than 2.5times higher forthe hydrocarboncase than the casewith no hydrocarbon layer.

Even when the reservoir width is only2 km, most of the hydrocarbon responsehas developed and is about two timeshigher than the direct wave background(no hydrocarbon layer).

The modeling shows that the refractedEM waves’ Ez component in a simplemodel with homogenous overburden is

influenced, but does not seem to be muchinfluenced by the variation in reservoirwidth if the width is 2 to 3 km when thereservoir depth is about 1 km.

More developmentA promising novel application of

CSEM, called Seabed Logging (SBL), fordeepsea exploration and prospect evalu-

3D FINITE DIFFERENCE TIME DOMAIN MODELING Fig. 8

1,000

500

1,500

1,000

500

500 1,000 1,500 2,000 2,500 3,000 3,500 4,000

500 1,000 1,500 2,000 2,500 3,000 3,500 4,000

1,500

2.5D model to study the effect of reservoir widthAxial section

Cross-sectionX-distance, m

X-distance, m

Z-d

ista

nce

, mZ

-dis

tan

ce, m

Reservoir, 1,000 m below seabed, thickness 120 m, resistivity 50 ohm-m

Sea water, resistivity 0.3 ohm-m Overburden, resistivity 1.5 ohm-m

MODELING RESULTS FOR DIFFERENT RESERVOIR WIDTHS Fig. 9

0

50

100

150

200

250

300

350

450

400

0

50

100

150

200

250

300

350

450

500

400

3 km

2 km

0

50

100

150

200

250

300

350

400

1 km

Traces of normalized receiver signals for source-receiver offset 0-4,200 m. Black traces = hydrocarbon-filled reservoir; red traces = water-filled reservoir.

New Views of the Subsurface

E X P L O R A T I O N & D E V E L O P M E N T

ation, has been presented.Even when restricted to available

technology not tailored for this purpose,SBL can aid exploration and prospectevaluation in a number of today’sdeepsea areas.

Dueto thepotentialof savingcosts by

reducing the number of dry explorationwells and through promotion of earlycash flow earned by an early discovery,the method would naturally get businessfocus in the future.With the developmenthistory of the seismic industry in mind,one can easily imagine the developmentpotential of the SBL technique. ✦

References1. Eidesmo,T., Ellingsrud, S., Kong,

F.N., Westerdahl, H., and Johansen, S.,“Method and apparatus for determin-ing the nature of subterranean reser-voirs,” Patent application number WO00/13046, 2000, filed August 1998.

2. Ellingsrud, S., Eidesmo,T., Kong,F.N., and Weterdahl, H., “Reservoiridentification using refracted wave,”Patent application number WO

01/57555, 2001, filed February 2000.3. Eidesmo,T., Ellingsrud, S., Mac-

Gregor, L.M., Constable, S., Sinha, M.C.,Westerdahl, H., and Kong, F.N., “Re-mote detection of hydrocarbon filledlayers using marine controlled sourceelectromagnetic sounding,” paper sub-mitted to EAGE 64th Conference & Ex-hibition, Florence, Italy, May 27-30,2002.

4. Sinha, M.C., Patel, P.O., Unsworth,M.J., Owen,T.R.E., and MacCormac,M.R.U., “An active source electromag-netic system for marine use,” MarineGeophys. Res.,Vol. 12, 1990.

5. Chave, A.D., Constable, S.C., andEdwards, N., “Electrical explorationmethods for the seafloor,” in Nabighi-an, M., ed., “Electromagnetic Methodsin Applied Geophysics,”Vol. 2, SEG,Tul-sa, Okla., 1991, pp. 931-966.

6. King, R.W.P., “Antenna in materi-al media near boundaries with appli-cation to communication and geo-physical exploration, Part 1: The baremetal dipole and Part 2: The terminat-ed insulated antenna,” IEEE Trans., An-tennas and Propagat., Vol. Ap-34, No.4, April 1986.

7. Kong, J.A., “Electromagnetic WaveTheory,” John Wiley & Sons, 1986

The authorsF.N. Kong (fk@ngi.no) hasbeen working at NorwegianGeotechnical Institute since1989. His research interestsinclude wave propagation,ground penetrating radar, an-tenna design, etc. He received aBSc in radio engineering fromXidian University and PhD in electronics in Cam-bridge University. In 1986-88 he was a post-doctoral fellow of the Royal Norwegian Councilfor Scientific and Technological Research.

Harald Westerdahl(hw@ngi.no) worked withNorsk Hydro from 1984 to1988 with data acquisitionand processing of well seismicdata. Since 1988 he hasworked at the Norwegian Geot-echnical Institute, mainly withspecial seismic and

georadar/EM R&D projects for the petroleum in-dustry. He has an MSc in petroleum geology andgeophysics from the University of Trondheim.

Svein Ellingsrud(sve@emgs.no) worked at Sin-tef/Allforsk in Trondheim as aresearch scientist in 1990-92and at Statoil R&D Centrewithin the area of petro-physics/geophysics in 1992-2002. He is vice president ofElectromagnetic Geoservices AS.He has an MSc and PhD in physics from theUniversity of Trondheim.

Terje Eidesmo (tee@emgs.no)started as a petrophysicist atStatoil’s research center in1991. He was project managerfor several research projects andbecame a director in 2000-02. He is a petrophysics profes-sor at NTNU. Now president ofEMGS AS, he has a PhD in

physics from the University of Trondheim.

Ståle Emil Johansen(sj@emgs.no) has 15 years’experience in the oil industry.He started his career in EssoNorge in 1985 and continuedin Statoil. He has worked with-in exploration and research andwhile employed with EMGS isalso a part time professor atUniversity of Science and Technology. He has a PhDin geophysics from the University of Oslo.

3D MODELING WITH RESERVOIR PRESENT1 Fig. 10

Longitudinal section (along long axis) Cross-section

1Snapshots of the EM field at three different instances (top, middle, bottom).

New Views of the Subsurface