GEOPHYSICS, VOL. 56, NO.2 (FEBRUARY 1991); P. 255-264. 19 FlGS.

Controlled-source audiofrequency magnetotelluric responses of three-dimensional bodies

N. B. Boschetto* and G. W. Hohmannt

ABSTRACT lower in frequency. However, as the size of the conduc­Modeling the controlled-source audiofrequency tor and its secondary-field response increases, large

magnetotelluric (CSAMT) responses of simple three­ transition-zone responses distort the data. dimensional (3-0) structures due to a grounded elec­ For both a plane-wave source and a finite source, tric bipole confirms that the CSAMT technique accu­ current channeling into a 3-D conductor from conduc­rately simulates plane-wave results in the far-field tive overburden enhances the response of a target. The zone of the transmitter. However, at receiver sites modeled response of a dike-like conductor shows no located above large conductive or resistive bodies, the better results for either the broadside or collinear presence of the inhomogeneity extends or reduces, configuration. The location and extent of such a body respectively, the frequency range of the far-field zone. are better defined when measuring the electric field Measurements made on the surface beyond a large 3-D perpendicular to the strike of the prism, but resistivity body display a small but spatially extensive effect due estimates are better when using the electric field to decay of the artificial primary field. parallel to the strike of the prism, irrespective of

Situating a 3-D inhomogeneity beneath the source transmitter orientation. Models designed from data permits an evaluation of "source overprint" effects. collected at Marionoak, Tasmania, yield results which When such a body is resistive, a slight shift in the indicate that the thin, vertical graphitic unit inter­near-fieldresponse to higher frequencies occurs. When a sected by drilling is detectable by the CSAMT method, body below the transmitter is conductive, it is possible to but probably is not the sole source of the large make far-fieldmeasurements closer to the transmitter or anomaly seen in the CSAMT data.

INTRODUCTION the far-field of the transmitter is known. Strictly speaking, plane-wave behavior can be determined only when the true

The controlled-source audiofrequency magnetotelluric geoelectric structure is known or the earth is one-dimen­(CSAMT) method is an efficient electromagnetic (EM) tech­ sional (1~D). Questions remain on applying the plane-wave nique widely used for mapping shallow resistivity structure approximation when the source waves must pass through an in the earth. By employing a transmitter (usually a grounded inhomogeneous, three-dimensional (3-D) earth. electric bipole) to generate source fields, the CSAMT The goals of this model study are to illuminate some of the method overcomes problems associated with weak natural effects that 3-D inhomogeneities have on CSAMT data, as EM fields used in conventional audiofrequency magnetotel­ well as to illustrate how the response of a 3-D body is luric (AMT) surveys. Although a bipole transmitter creates affected by a finite source. fields which originate a finite distance from receiving sites,

THE CSAMT METHOD measurements made in a far-field zone of the source behave like those of an infinitely distant plane-wave source. Thus, Previous work simple plane-wave apparent resistivity calculations are fre­quently applied in CSAMT data analysis. Since Goldstein and Strangway (1975) first introduced the

The plane-wave assumption is only reliable if the extent of application of an active source for collecting magnetotelluric

Manuscript received by the Editor November 7, 1989; revised manuscript received August 6, 1990. "Formerly Department of Geology and Geophysics, University of Utah; presently ERM-Southwest, Inc., 16000 Memorial Drive, Ste. 200, Houston, TX 77079. +Department of Geology and Geophysics, University of Utah, Salt Lake City, UT 84112. © 1991 Society of Exploration Geophysicists. All rightsreserved.

255

256 Boschetto and Hohmann

(MT) data in the audiofrequency range. the CSAMT method has been employed in a variety of exploration environments. Goldstein and Strangway (1975) tested their technique in the field over a massive sulfide body after performing scale­model studies. A geothermal area was the target of Sandberg and Hohmann (1982), who also showed with numerical modeling that the method approximates plane-wave results for a 3-D inhomogeneity located a distance greater than five skin depths from a bipole transmitter. Ostrander et a1. (1983) mapped an electrical anomaly over an oil field. Another application of the CSAMT technique to petroleum explora­tion is structural mapping, as done by Hughes and Carlson (1987). Bartel and Jacobson (1987) determined the depth and electrical properties of a volcanogenic thermal anomaly using CSAMT data which incorporated a correction factor in the results from near-field frequencies.

Thorough overviews of the CSAMT method-history, theory, applications, field techniques, data interpretation, case histories-have been given by Zonge et al. (1986) and by Zonge and Hughes (1988).

Considerations for a bipole source

Because the CSAMT method uses an artificial source. factors beyond those for the AMT technique (Strangeway et al., 1973) concerning data collection and analysis must be taken into account. In this study, a grounded electric bipole source is simulated to address some of these factors.

Figure I schematically illustrates in plan view the areal limitations of a CSAMT field survey. The configuration shown is for scalar measurements in which only one orthog­onal pair of the electric and magnetic field components is measured. All results presented use the component of the electric field parallel to the transmitter tE; in Figure t) in either the broadside or collinear configuration. The stippled areas in Figure 1 denote far-field zones where the fields behave approximately as plane-wave fields and are large enough to be measured accurately. Within these regions, the simple plane-wave, or Cagniard (Cagniard, 1953), apparent resistivity relationship

Pa = (J)~ (lEI) 2

wj..L IHI can be reliably applied, where j..L is the magnetic permeability offree space, w is the angular frequency. and E and H are the orthogonal components of the electric and magnetic fields, respectively. The phase difference

<I> == <P£ - <PH (2)

is also calculated, where cPE and <f>H are the phases of the electric and magnetic fields relative to the transmitted signal. In the far field, this quantity is equivalent to the plane-wave phase. As determined by Sandberg and Hohmann (1982), the far field of a bipole transmitter in a homogeneous earth begins at a distance of approximately three to four skin depths for the broadside configuration and five skin depths for the collinear configuration. A skin depth is defined as

0==503 (3)\j.rI~ m'

where p is the resistivity in n . m of the medium and I is the frequency in Hz. Hence, for reliable application of MT data reduction. one must measure the fields at distances at least as great as those described above,

In the un stippled areas in Figure I, the fields are either too small to be measured accurately (beyond and between the stippled zones) or do not behave as plane waves (near the transmitter). The region between the far-field zone and the transmitter is commonly divided into two zones accord­ing to distance from the source r: the near-field zone (r ~ 0) and the transition zone (0 ~ r ~ 3 - 4 or 50). In the near-field zone, the fields are frequency independent, or "saturated;" and the impedance is proportional to I/r (Zonge and Hughes. 1988). Application of plane-wave apparent resistivity calcu­lations to data collected in the near field of the transmitter yields apparent resistivity values which increase linearly as frequency decreases on a log-log plot. This behavior is termed the "near-field rise" and occurs because the plane­wave apparent resistivity requation (l)] is proportional to 1iI in the near-field zone where E and H are constant. The phase difference [equation (2)] tends toward zero in the near field. The transition zone is aptly named; the behavior of the fields within this zone are gradational from the near-field response to that of the far field.

A~ALYSlS OF 3-D N1JMERICAL RESULTS

The models in this study are designed to represent se­lected situations commonly encountered in the field. The computer-intensive nature of 3-D modeling limits the com­plexity of model geometries, but even our simplified models shed new light on CSAMT survey planning and data inter-

Ey

approx. scale

" 0 1000m.. L..........J

Collinear

FIG. 1. Plan view schematic of scalar CSAMT field config­uration. Transmitter parallel to y-axis is labeled T't:. Stippled areas indicate far-field measurement zones (after Longe and Hughes, 1988).

r"

---

257 CSAMT Response of 3-D Bodies

pretation. The models include a large buried prism located either some distance away from the source. within the modeled CSAMT survey area, or directly beneath the trans­mitter. The effect of an inhomogeneity beneath the transmit­ter on the response of a 3-D target in the survey area is also examined. A 3-D model with conductive overburden illus­trates the effects of current channeling. Thin, vertical prism models evaluate the detectability of dike-like conductors with the CSAMT method.

We used the volume integral-equation program of New­man et al. (] 986) to compute the electric and magnetic fields due to a 1000 m bipole transmitter at point receiver sites for ]-D and 3-D models. From the field values, we calculated the apparent resistivity and phase difference according to equa­tions (1) and (2). A frequency range of I to 1000 Hz proved adequate for all models except those with an inhomogeneity located beneath the transmitter, for which the frequency range was extended downward to 0.01 Hz. For comparison, we calculated the MT responses of the 3-D models using the algorithm of Wannamaker et al. 0984a). All model results were computed on a VAX 11/785 computer.

Large buried prism

Figure 2 shows the model configuration for a large buried prism located within an otherwise homogeneous earth. Two cases are considered: one in which the body is conductive relative to the host medium and the other in which the hody is relatively resistive. In both cases, the two vertical planes of symmetry were used to advantage (Tripp and Hohmann, ]984); we discretized one quadrant of the body into 64 250 m cubic cells. This model requires roughly 20 minutes of CPU time per frequency.

Figures 3 and 5 display 3-D CSAMT apparent resistivity and phase sounding curves for the large conductor and resistor models, together with the plane-wave response (3-D MT) and the 100 n'm half-space data (I-D CSAMT). Arrows in Figure 2 show the two sounding positions. Note that x = oabove the center of the body. All presented results are for the broadside configuration in which the transmitter, Tv' is parallel to the y-axis and calculated apparent resistivities are proportional to E.,. /Il, .

First consider' the response of the conductive prism. Figure 3a shows marked similarity between the controlled­source and the plane-wave responses at x = 0 for the conductive body. This agreement continues well into the near-field zone of the host, as indicated by the I-D CSAMT apparent resistivity curve. The presence of the large conduc­tor below the receiver seems to extend the far-field fre­quency range to that of a medium with a smaller skin depth. However, the departure of the 3-D CSAMT phase from that of the 3-D MT case at 1 Hz indicates that the transition into the near field is imminent.

Comparing the sounding curves of Figure 3b shows that at the receiver location beyond the conductor, the 3-D CSAMT apparent resistivity curve at lower frequencies exhibits values somewhat reduced from what is expected from the MT results. Kuznetzov (1982) and Zonge and Hughes (1988) observed a similar phenomenon in scale and 2-D numerical model studies and termed such behavior the "shadow ef­feet. " Because the transmitted fields decay from the source

to the receiver, the secondary field of the body comprises a greater portion of the total field beyond the body. Figure 4 illustrates the areal extent of the shadow effect for the buried conductor at a frequency of 10 Hz. The plan-view plots cover the entire 6 x 6 km survey area. Beyond the conduc­tor, both apparent resistivity and phase anomalies noticeably extend away from the transmitting bipole. Though small, the shadow effect persists to a distance of 6000 m along the

PLAN r ~Figures 30.50 ~ f:,--F igures 3b. 5b~

TX1l< II " *I II t " ~ I

t lC " lC_X

$X""iXllXX

I IL.. ....-J

" x v " -2000rn­x X lC

" " x

X II

2 R . eceivers

lC II

" II )(

~ y

r, Ty

100 n'l\

! z

FIG. 2. Model configuration for a large buried prism located within a CSAMT survey area. Transmitter parallel to x- or y-axis is labeled T, or T}"" respectively. Receiver spacing is 500 m. Arrows indicate locations of soundings in Figures 3 and 5.

103 103

1E 10 ~ 102

s E

s <>?' 101 ­~101

- IDCSAIH

100 L,.!__...l..-__..L-_--l 10°1 I

CIJ, 9:t:=;=:-;--: : ~M90~ -e?' 0 .... .......

-s. 100 10 1 102 '0 3 100 10' 102 10 3

f,equency (H.) FreQuency (Hz)

(0 ) (b)

FIG. 3. Soundings (a) above (at x = 0 m) and (b) beyond large 1 n·m conductor at x = 1500 m.

258 Boschetto and Hohmann

x-axis from the center of the 3-D body, which is three times the width of the body and six times its depth extent (results not shown).

For the collinear configuration (using transmitter T, in Figure 2), a larger shadow effect in the opposite sense occurs (not shown). Because the secondary electric field of the conductor adds to the primary field along the x-axis beyond the body, the collinear 3-D CSAMT apparent resistivity increases. This apparent resistivity increase is greater than the corresponding decrease in the broadside case because of the larger rate of decay of the primary fields with increasing collinear distance from the transmitter.

Examining the sounding curves at the same receiver sites for the large resistive body in Figure 2, Figure 5a shows that at the receiver at x == 0, the resistive body has an effect on the modeled data which is opposite to that of the conductive body. The large skin depth in the resistor causes the transi­tion to near-field behavior to occur at a higher frequency than for the 100 n'm half-space. While it is difficult to separate the response of the body from transition-zone effects at frequencies greater than 100 Hz, the lowest possi­ble frequency at which the transition from the former to the latter occurs in the 3-D CSAMT data is approximately 80 Hz. However, the accompanying 1-0 CSAMT curves in Figure 5a show that this transition in the host begins at approximately 50 Hz.

At the receiver beyond the body away from the source (Figure Sbl, the shadow effect is evident, but in the contrary sense and to a slightly greater degree than in the case of the conductor; 3-D CSAMT apparent resistivity values at lower frequencies are larger than both the corresponding 3-D MT and 1-0 CSAMT values. Galvanic effects in the resistor generate a secondary electric field which adds to the primary field along the x-axis. Because the primary field of the transmitter decays in the modeled earth, this secondary field contributes more to the total-field response in the CSAMT case than in the MT case. Hence, the 3-0 CSAMT apparent resistivities are greater than expected. Again, it is difficult to distinguish to what degree the various effects contribute individually to the observed results, but this apparent resis­

38

(a) (b)

FIG. 4. Plan-view (a) apparent resistivity and (b) phase plots at 10 Hz over the entire survey area above the] n'm body in Figure 2. Symmetry across the x-axis is used, and con­ductor location is shown by dashed lines. Transmitter is labeled Ty- Hachures mark zones of minimum measured signal strength. Stipples denote approximate extent of the shadow zone. Distances of three skin depths in the 100n· m host from the center of the bipole are indicated with arrows.

tivity enhancement may also be abetted by an increase in frequency of the inception of the near-field rise caused by the intervening resistive body (Figure 5a). For the conductive body, the effect continues on the x-axis to at least 6000 m from the center of the body (results not shown).

In the collinear configuration (not shown), a correspond­ing reduction in 3-D CSAMT apparent resistivity values occurs, but the magnitude of the effect is not appreciably greater than for the broadside case. Even though a larger effect is expected because the primary collinear fields decay more rapidly away from the source than broadside fields, it appears that the higher frequency. near-field effects caused by the resistor counteract part of the shadow effect.

Transmitter overprint

CSAMT soundings sometimes vary significantly with transmitter placement, even if the transmitter-receiver sep­aration is held constant. This troublesome phenomenon is called transmitter overprint and appears to be a function of local geoelectric structure beneath the transmitter. To inves­tigate such overprint, we computed several 3-D models with inhomogeneities located directly below the bipole source. Figure 6 displays the model configurations. First, we present several examples in which the body beneath the source (body A, C, or C' in Figure 6) is the only one in the half-space. Second, we address the effect that an overprint caused by a body beneath the source has on the response of a 3-D target (body B in Figure 6).

Model discretization proceeded according to the dictates of Ting and Hohmann (1981) and Newman et al. (1986). Discretization required to achieve accurate results varied for each model. The 10000 n'm body A and the I fl-rn body C were discretized into 250 m cubic cells. Finer discretization was necessary for the I n'm bodies C' and A; body C' was divided uniformly into 125 m cubic cells, while 1 n· m body A was discretized into 125 m cubic cells in the upper 250 m and 250 m cubic cells in the lower portion. The latter model, which contains the largest number of cells, requires 2.2 hours of CPU time per frequency.

103 103

2E 10 ~

~ ~

~101 o-c 30 C$A"T Q.."

101

10° ' rio 90 "

~ob ;::;::­-------J...~ oJ

10° 101 10 2 103 10' 10 2 10 3

Frequency (Hz) Frequency (Hz)

(0) ( b)

FIG. 5. Soundings (a) above, at x = 0 m, and (b) beyond large 10 000 n'm resistor at x == 1500 m. Compare with Figure 3.

259 CSAMT Response of 3·0 Bodies

Figure 7 contains five sets of sounding curves from the broadside location shown in Figure 6. The solid curve without symbols is the 100n'm half-space response, shown for reference. The other four curves are results with various 3-D bodies located beneath the transmitter but without the target (body B). In all cases the far-field data; i.e., for frequencies greater than 100 Hz converge, to half-space results. However, the lower frequency data, as the receiver site shifts into the transition and near-field zones, depart from the I-D response. The data are most perturbed when the large I n'm body A is below the source. This conductive body sets up a large secondary field as well as suppresses the transmitted fields. However, when the same size body is present as a 10 000 n· m resistive unit, only a small shift in the near-field rise to higher frequencies occurs. Zonge et al. (1986) have observed similar phenomena in field tests.

Some practitioners of the CSAMT method claim that placing a transmitter above a known conductive unit extends the far-field zone, closer to the transmitter or lower in frequency, due to the smaller skin depth in a conductor. Obviously, with a conductor similar to body A of Figure 6, the transition-zone distortions would outweigh the benefits of shifting the near-field rise to lower frequencies. However, when a smaller conductor of the same resistivity (body C or C) is located beneath the transmitter, the results in Figure 7 indicate that an extension of the far-field zone without much

PLAN r---- ----, SOOm Broadside SoundingI ~ and C A: ~ Location,I '-.;" -j--i I i-8:~ I I I I I 10

1()(X~X~O)( ·-x: :t, : i l( II( x )( 'X x O )( )(

, I I I ­I '•• - -- -. • x )( )( x-x-x-*- x )( I I XXXXXXXX)(

L.... ,,__ •__ ~ )( )( )( x x x )( )( )( 1---2000m------1

~ 5 Receivers

"-- Collinear Sounding x Location

I y

CROSS-Sf C nON -x

, 0'

! o lOOOm \-..---J10.000 SCALE.Q.,.,

distortion may be possible. The larger secondary fields at the surface created by the shallower body C' cause greater expansion of the far-field zone and a little more transition­zone fluctuation than when the smaller conductor is more deeply buried (body C).

A more practical test of these transmitter overprint models is to examine their effect on the CSAMT response of a 3-D target. In this study, body B in Figure 6 represents a 10n'm conductive target.

First consider the case in which both the large I n· m conductor A and the lO n'm target B are in the otherwise homogeneous 100 n ' m host. The question to be answered is whether or not the large transition-zone fluctuations seen in Figure 7 obscure the signature of the target B. For reference, Figure 8 shows the response of the target alone, with no inhomogeneity below the transmitter, in pseudosec­tion format. The pseudosections are through the center of the target, coincident with the x-axis. The 10n· m conductor B has a well-defined response (including a small shadow zone) which pinches out and yields to source effects at lower frequencies. Comparing the results in Figure 8 to the plane­wave response of the same model (Butterworth, 1988) indi­cates that the CSAMT response of target B is representative of the MT results down to a frequency of about 30 to 70 Hz, where the transition from the far field to the near field begins.

To illustrate the effect of a large conductive body beneath the transmitter, Figure 9 shows pseudosection results when both bodies A and B are present. Comparing Figures 8 and 9 shows that the strong response of conductor A beneath the transmitter reaches well into the region of the target re­sponse at lower frequencies. Thus, if regarded indepen­dently from the data near the transmitter. both apparent resistivity and phase data below the smaller target B would erroneously indicate a deeply extending conductive target. Apparent resistivities which decrease with frequency in the target response and show no sign of near-field effects,

10,000 ," "

1000

s ~

I I 01 10 100

I

1 C

j 01 10 10 100

z F,eQuency (Hz)

FIG. 6. Transmitter overprint models. Various 3-D body combinations are described in the text. Transmitter is la­ FIG. 7. Apparent resistivity and phase soundings for different beled Ty • Receiver spacing is 500 m except above body B model geometries at broadside location on x-axis shown in where it is 250 m. Figure 6. The target B is not present.

260 Boschetto and Hohmann

coupled with high phase values at low frequencies (Wanna­maker et aI., I984b), would support this conclusion. To verify that this overprint by conductor A is an artifact of an artificial source field, we computed the plane-wave response of the two bodies (Butterworth, ]988). The MT signature of the larger conductor A does not extend into and alter the anomaly due to target B, as when the bipole source is used.

Next, we investigated the possibility that the small, shal­low (50 m deep), ] n'm body C' below the transmitter actuaUy lowers the frequency of the far- to near-field transi­tion in the response of the 3-D target B. The response of conductor C' is not quite as strong at low frequencies as that of the larger conductor A and, indeed, the measurements with the conductor C' present under the transmitter are representative of the desired plane-wave response of the target down to a frequency of about] to 3 Hz (Butterworth, ]988). Thus, the far-field measurements made above the target are extended downward by at least a decade in frequency by the overprint of body C'. Although this model produces desirable results, we do not recommend the prac­tice of situating a transmitter above a conductive unit because of the possibility of large overprint problems as illustrated in Figure 9.

Before leaving the transmitter overprint models, we com­ment on the collinear configuration results. Soundings at the location on the y-axis in Figure 6 with the 1 n·m body A or

Ion. f .. ,." Ta'llI"rB 'oD"

Pu cn.) "' •• (oe;;)

sooo 6000~X T~ 1000 looa moo 4000 500(· 6000. 10:>0; ( .. _-)(

T, 1000 2000 [~~:~o ----l,­lOOO~ s 300 \ q~~

:::1'1)\ \0~ IOO~, J~~ l 'b V,A

~ so f~,\-'b '" LJ~/O,5~~ ::]\ ~ :: "~\ -. \rJ Vf\ '-:,.'~ -, "I::; I V '\J- <;

\ \ ,LR'''' <,,'\ ('JOe,_

(a) (b)

FIG. 8. (a) Apparent resistivity and (b) phase pseudosections coincident with the x-axis through center of 10 n'm target B shown in Figure 6. No anomalous 3-D body is present beneath the transmitter. Transmitter T, is in upper left corner of each plot. Stipples indicate approximate extent of the shadow zone.

Ill. 1012. &o,,~ A 'ody A ro.~.r i 9w-(ol.g)

IU·.

,OOC! ' 1000 2QCC' )000

locl~~~'_~oo~x IJ'~ li~ I

c ~ \) , j JO~7() ,Iv - I ~ : ,_III 't -~

~~ (bJCol

FIG. 9. (a) Apparent resistivity and (b) phase pseudosections coincident with the x-axis through the center of a large 1 n'm conductor A and 10 n· m target B shown in Figure 6. Transmitter Ty is in upper left corner of each plot. Compare with Figure 8.

C' beneath the source do not display the transition-zone fluctuations of Figure 7; rather, because the secondary electric fields for the collinear configuration contribute to, rather than oppose, the decaying primary fields along the y-axis, the response in transition-zone frequencies is very close to the plane-wave response of the 100 n'm host (Butterworth, 1988). The presence of conductive body C' or A extends far-field results lower in frequency by one or two decades, respectively. While the collinear configuration is not often employed because of decreased field strength, under certain conditions it might be possible to use the above information to advantage.

Conductive overburden

Newman et al. (1986) have shown that the transient EM response of a 3-D conductor is enhanced when in contact with conductive overburden because current channels from the overburden into the more conductive body. Figure 10 illustrates the geometry of a model designed to investigate this phenomenon for the CSAMT case. The body was discretized into 16 125m cubic cells per quadrant; computing its response requires 5 minutes of CPU time per frequency.

Figures 1] and 12 show pseudoseclions along the x-ax.is, through the center of the 3-D body.. and depict the model response with and without the overburden, respectively. Two major differences can be noted between the different responses. First, the smaller skin depth in the conductive overburden causes the anomaly of the 3-D body to develop at lower frequencies than when the body is in the half-space. Second. the anomaly is stronger when the overburden is present. The apparent resistivity is more representative of the true resistivity of the body, and the phase anomaly, while not larger in an absolute sense, is larger when contrasted with the phase response of the host. The stronger 3-D anomaly in Figure 1] indicates that current is channeled from the overburden into the attached body. This augmen­tation of the response appears to be independent of the artificial source because plane-wave results (not shown) are very similar to those in Figure II.

We tested another model in which the overburden was thinned to 20 m and hence disconnected from the body. The 3-D response (not shown) is intermediate between the above two results, suggesting that although detached from the overburden, the large 3-D conductor still channels current from the thinner layer above it.

The collinear configuration produces results comparable to the above broadside cases.

Thin vertical prism

Here we examine the CSAMT response of a thin, vertical, conducting prism, approximating a steeply dipping mineral­ized dike or fracture zone. Figure 13 illustrates the model configuration. Without utilizing symmetry, we discretized the body into 90 prismatic cells; the computation requires approximately 1.25 hours per frequency. This model com­pares the responses of the dike-like structure for the four transmitter orientations shown in Figure 13 and evaluates the detectability of the prism at two burial depths and for the two different modes of ex.citation.

261 CSAMT Response of 3·0 Bodies

Considering the CSAMT response of the more deeply buried 000 m deep) conductive prism in Figure 13, Figures 14 and 15 show pseudosection results using the broadside transmitters T.v I and Tx2 ' respectively. The pseudosections are through the center of the prism, perpendicular to its strike, and coincident with the x-axis. The yx-polarization plots of Figure· 14 show a strong and areally broad anomaly above the body. However, the body is barely detectable when the incident electric field is

PLAN 500", ~ r---, , I I I! I 10 x_xt- ] '"''~''~O'''

llXXXXJlOx)CI I ~

I( II( • )C-)f; -)£--*- )I til

~)(.x:x .. )(¥ •

• " It It " " " " "

~ 1 Receivers

! y

CROSS-SECTIONu; L. • M U It I •

II ...~ 1250",140m 1.0.· ..

100.0'111

o 1000m L--.--.J

SCALE

1 z

FIG. 10. Conductive overburden model configuration. Trans­mitter is labeled Ty • Receiver spacing is 500 m. except above the body where it is 250 m.

P.,.(,Q .• ) rf..,.("'.) t---" 2000 30~l('

I ...\ \ t'1·/ -,

'\. -, f'I

(0) (bJ

FIG. 11. (a) Apparent resistivity and (b) phase pseudosec­tions coincident with the x-axis through the center of the 3-D conductor in Figure 10. Overburden is 10 n·m.

perpendicular to the strike of the prism (Figure 15) and the anomaly could easily be obscured by other geoelectric struc­tures or static effects in a field situation. Corresponding plane­wave results (not shown) compare well with the CSAMT results down to frequencies in the 3 to to Hz range, when source effects become evident.

The CSAMT response of the shallow (30 m deep) prism using the two broadside transmitters TVI and Tx2 ' shown in Figure 13, is very similar to that of thedeeper prism except that the anomalies are larger in magnitude (Butterworth, 1988). Again, both polarizations are in close agreement with 3-D plane-wave results (not shown).

For the collinear configuration (results not shown), using transmitters TXI and TY2 in Figure 13, the overall appearance of the anomalies is like that of the corresponding (xy- or y x-polarization) broadside configurations.

Pyo(ll.1 t;:.d.:: ',. (,,~.l

r , 1000 :'000 3000.~co._)( 1000 ----)(

\ " 105 ~ \ I ~~ lOClll\ t~. ~ (~6'"

; '·1;1 \ \ ~ /",.r!! ~ \ 1'\ 1 ~ lOV/ \'I\U/1 ,(, 'f:~I() \ \ \\\.1 ('05

,~\\ '?t, \I! \\ -, 1',11\.:· Y~f\ :

,LE. (bl <oj

FIG. 12. (a) Apparent resistivity and (b) phase pseudosec­tions for the same model as Figure 11 without the overburden.

PLAN ~~h'2

~ Receivers -<-­

t-,

.1000 .. -x+rKI: a :T

-'11­25"

~OOrn

SCALE y

TK, Ty, CROSS-SEC TION -x - 300, 'OO~O"

lOon·. In.. 25"

z

FIG. 13. Model configuration for thin, vertical prism. Bipole transmitters parallel to x-axis are labeled Tx ; those parallel to y-axis are labeled Ty • Receiver spacing is 100 m above the prism and 500 m elsewhere. Two burial depths are consid­ered: 30 m and 100 m.

262 Boschetto and Hohmann

Field example: Marionoak

A deep conductor at the Marionoak property in Tasmania, Australia, has been the object of recent geophysical interest and controversy. Both transient electromagnetic (TEM) and CSAMT data indicate the presence of a large conductive body at depth (Eadie et aI., 1987). CSAMT pseudosections which usc the electric field parallel to the geologic strike and to the apparent strike of the body are shown in Figure 16. A borehole drilled into the interpreted anomalous zone, be­tween stations - 2 and - 5 in Figure 16, intersects a good conductor, but the width of the inhomogeneity as drilled is surprisingly small relative to its depth of burial. Using CSAMT data collected by Zonge Engineering and Research Organization and reports provided by Aberfoyle Resources Ltd., we designed a 3-D model which addresses the CSAMT detectability of a conductor similar to the one thought to exist at Marionoak.

The geologic setting at Marionoak is complex. Two groups of resistive, vertically dipping, thinly layered sedimentary rocks form the regional host. The deep conductive target, a graphitic zone roughly 150 m wide, was penetrated by drilling at a depth of approximately 600 m. The CSAMT anomaly associated with the graphite has a strike length of at least 1200 m.

Because the above geologic situation is not readily repre­

~ «, (n_) 1 ~,.(64, I

'::r~~'~ ~,"oo '000 r' "··T~\-I:\- '~l-' 100 1100 I:"~ Ij \j\''::)., ~ \ \ll JI\, \V,O~\ fil~1 ) '01_ "'"I _______

"et, ~ ~ 101.~~\~~~ If\~ , \~V' I ''6 -.

'It, ..?lOb I

I

"~;~j:~ "'--­00' ~ I

(0) (bl

FIG. 14. (a) Apparent resistivity and (b) phase pseudosec­tions coincident with the x-axis through the center of the deep vertical prism in Figure 13. Configuration is broadside and electric field measured parallel to prism strike. Trans­mitter I y l is in upper left corner of each plot.

"Jlk...

l ,o.,(a.l ~ •., '''.' -3000 -'2000 -1000 ~oooc 1000 woo.

, _¥ ~ogi:':"":o lCOO·1000:.:::::......::·~;::...""--:-0 . >~~._~1000~

I

::1 --"r"- J~ 100

/;;0, ________ ~ ::i;=:J'I !(1~~_~~ \ 1 Ie Il 0\ \

(0) (bl

FIG. 15. (a) Apparent resistivity and (b) phase pseudosec­tions coincident with the x-axis through the center of the deep vertical prism in Figure 13. Configuration is broadside and electric field measured perpendicular to prism strike. Transmitter TX2 is off the line of the pseudosection.

sented with a simple 3-D model, instead of simulating the entire setting at Marionoak, we examined the response of a similar target to determine whether it could produce an anomaly like that observed, Figure 17 illustrates the model configuration. The broader, moderately conductive 300 n'm unit laterally confines the deep, narrow 3 n'm target.

Figure ]8 shows apparent resistivity and phase pseudosec­tions through the center of the modeled inhomogeneity along the x-axis. We present only the results which use the electric field parallel to the strike of the target because we found that the detectability of the 3 !}. m prism is greatest for this mode. The 3-D body containing the thin conductive prism produces a broad, low-amplitude anomaly. Comparing the results in Figure 18 to the case where the 3 !}. m prism is replaced by 300 n'm material, leaving only a large 300 n'm body in the half-space (Figure 19), shows that the anomaly in Figure ]8, though significantly smaller than that seen in the Marionoak data, is due mainly to the 3 n'm prism. Indeed, the results in Figure] 9 are barely distinguishable from the half-space host

Po (.Q,·m)STATION -6 -5 -4 -3 -2 -1 0 1 2 3

4096 lin (, j , ((<un D ::>i. iJ iii

";:j 2048 ::I: 1024 >­ 512U Z 256 w :::> 128 o w 64 QI!. u.. 32

16" ,[) ) "" !I! II

(0)

ep (mrcd) STATION

-6 -5 -4 -3 -2 -1 0 1 2 3 4096 i <::L:> i < i' ........ '<QOiS!OJ I

~ 2048 ::I: 1024 >­ 512U Z 256UJ

:::> 128 o UJ 64 ~ LL. 32

16' ) ! ( { • YV> -! ! ( ,

( b)

FIG. 16. (a) Apparent resistivity and (b) phase field data from Marionoak, Tasmania. Electric field is parallel to geologic strike. Station spacing and receiver dipole length are 100 m.

263 CSAMT Response of 3-D Bodies

response (not shown). The thin, 3 n'm prism, when modeled along in the 1000n'm half-space (results not shown), yields a similar (slightly smaller) anomaly to that in Figure 18.

The weak response of the modeled prism indicates that factors other than the thin graphitic zone intersected by drilling influence the large anomaly in the Marionoak CSAMT data. Possibilities include a graphite body which is substantially wider than 150 m where not drilled or which laterally splays off the main mineralized zone into multiple graphite lenses. The hosting units may also be more conduc­tive than expected. Additionally, low apparent resistivity values could be caused by current channeling in the verti­cally dipping host beds.

CONCLUSIONS

The results of this study confirm that the CSAMT method produces results similar to plane-wave responses for 3-D targets in the far field of a bipole transmitter. However, a large conductive or resistive body below a receiving site extends or contracts, respectively. the far-field frequency range.

Measurements on the side of a large 3-0 body away from the source show that the primary-field decay has a small but areally extensive effect on CSAMT data. As the strength of the artificial primary field decreases away from the transmit­ter, the secondary fields of the 3-D body contribute more to the total-field response. Shifting of the far field caused by a

PLAN

r--750m--l

TV]rt-''I( )(

I--Y/.T--I I ~ I

i ~ : I F;/;:I I

'I()(i")(l~)('I()(I)(X

Ie

x g x -)(

N

x'-6000m

I ~ I I 88 I I V/:;i I L_~_-l 1

i y

o 300m L..--J SCALE

CROSS-SECTION

~e.---,1(,-6000", e

8 ~Rece;vers _)(

+If-pon.!~ iooo n,

3 n'm

+z

FIG. 17. Marionoak model configuration. The WOO m bipole transmitter, labeled Ty , is 6000 m from the center of the 3-D body. Receiver spacing is 75 m above the inhomogeneity.

large inhomogeneity also affects CSAMT measurements made beyond the body. For the broadside configuration, apparent resistivity values are slightly depressed or en­hanced beyond a conductor or resistor, respectively. The opposite is true for the collinear case.

When an inhomogeneity is located beneath the transmit­ter, the effects on broadside measurements made in the modeled CSAMT survey area range from minor to consid­erable. When the body below the source is resistive, a slight shift of the near-field zone to higher frequencies occurs. A smaller 3-D conductor below the source actually appears to extend the far-field zone to lower frequencies or to locations nearer the transmitter. However. as the size of the conduc­tive body and the secondary-field responses increase. appar­ent resistivity values at transition and near-field frequencies are greatly depressed. Soundings from the collinear config­uration when a 3-D conductor lies below the transmitter do not display the large transition-zone response seen in the broadside results; this contributes to a significant increase in the low-frequency range of the far-field zone.

Current channeling from conductive overburden into a 3-D conductor enhances the response of the target relative to

3 n... To,,,e, -It ­

300 .a ... Condu~'"r Pyx (n·.)f-o------1 ·300 0 300

2048

1024

r 512

> v Z 256 (0) ::J 0 128

.....

~ =:==d

UJ

0<

64

~100~,~ 32~

16

3Jl.. Ta,g,,' --I I­

300n·.. Co.dlJ~'Of <PYX (deg) t--.-.-~ ---1

-600 ·300 0 300 600 ~~.!.-X-900 1 Iii4096-1~200

2048

1024

X ~ 512 >­~ 256 L 'Vv ....

(b) a 12B :;I C:J.... ... ...

64

:.10 32

16 1 --- - J

FIG. 18. (a) Apparent resistivity and (b) phase pseudosec­tions coincident with the x-axis through the center of the 3-D target in Figure 17. Transmitter is located at .r --'- -6000 m and is parallel to the y-axis.

264 Boschetto and Hohmann

the situation without the overburden. The effect is indepen­dent of the artificial source.

Model results indicate that no preference exists for either broadside or collinear configuration when the target is a dike-like structure. The location and extent of such a con­ductive prism are better defined when the transmitter is oriented perpendicularly to body strike. However, apparent resistivity values are more representative and depth of exploration is greater when the transmitter is parallel to the prism's strike. A prism buried four times its narrowest horizontal width in a host 100 times more resistive is at the limits of CSAMT detection when measuring the electric field across the strike of the body.

Modeling the response of a conductor similar to the one found at Marionoak, Tasmania, suggests that the deep, thin inhomogeneity intersected by drilling is likely detectable using the CSAMT method, but the large anomaly seen in the field data is probably not entirely attributable to the thin

300n·,. ConJ.'tor PYX <n·.) '" ·1-300 0 300 600 900 1200.!'.--)(

; Iii iJ

.. :I:

>­u z I

--~--(0)

w 256f--­::> 0 128 '000 .... '" ....

64

321..70 0 0 16

epyX (dellJ

·1200 ·900 ·600 ·300 0 ;,vv ..,...,v >"'''' ._-­4096. • t ii' i I

2048

x >­ 44 U Z.... (b)::l o W At .....

-';0

32

16 Ll -'

FIG. 19. (a) Apparent resistivity and (b) phase pseudosec­tions for the same model as in Figure ]8 except that the thin. 3 n'm prism is replaced with 300 n'm material.

conductor at depth. Surrounding the modeled thin prism with a more conductive unit does not significantly enhance the response of the prism.

ACKNOWLEDGMENTS

This project was funded by a consortium of corporations including Amoco Production Co., ARCa Oil and Gas Co., Chevron Resources Co., CRA Exploration Pty. Ltd., Stan­dard Oil Co., and Unocal Corp. We extend our gratitude to C. M. Swift who inspired and guided this study. to Aberfoyle Resources Ltd. for contributing the field data, and to the Department of Physics at the University of Utah for use of its VAX 111785 computer.

REFERENCES

Bartel, L. C.; and Jacobson. R. D.; 1987, Results of a controlled­source audiofrequency magnetotelluric survey at the Puhimau thermal area. Kilauea Volcano. Hawaii: Geophysics, 52. 665-677.

Butterworth, N. A., 1988, Controlled-source audio-frequency mag­netotelluric responses of three-dimensional bodies: M.Sc. thesis, Univ. of Utah.

Cagniard. L., 1953. Basic theory of the magneto-telluric method of geophysical prospecting: Geophysics. 18, 605-635.

Eadie. E. T .• Silic, J.. and Hungerford. N.. 1987. Marionoak-a very deep conductive target: Presented at the 5th Ann. Conf .. Austral. Soc. Expl. Geophys.

Goldstein. M. A., and Strangway , D. W., 1975. Audio-frequency magnetotellurics with a grounded electric dipole source: Geo­physics. 40. 669-683. .

Hughes, L. J .. and Carlson N. R., 1987. Structure mapping at Trap Spring Oilfield. Nevada. using controlled-source magnetotellu­rics: First Break. S, 403-418.

Kuznetzov, A. N.. 1982, Distorting effects during electromagnetic sounding of horizontally non-uniform media using an artificial field source: Izvestiya, Earth Physics. 18, 130-137.

Newman. G. A., Hohmann, G. W., and Anderson, W. L., 1986, Transient electromagnetic response of a three-dimensional body in a layered earth: Geophysics. 51. 1608-1627.

Ostrander. A. G., Carlson, N. R.• and Zonge, K. L., 1983. Further evidence of electrical anomalies over hydrocarbon accumulations using CSAMT: 53rd Ann. Internal. Mtg., Soc. Explor. Geophys .• Expanded Abstracts, 6(}...{i3.

Sandberg. S. K.• and Hohmann, G. W., 1982. Controlled-source audiomagnetotellurics in geothermal exploration: Geophysics. 47, 100-116.

Strangway, D. W.. Swift. C. M., Jr., and Holmer. R. C.; 1973. The application of audio-frequency magnetotellurics (AMT) to mineral exploration: Geophysics, 38, 1159-1175.

Ting, S. c.,and Hohmann, G. W., 1981, Integral equation modeling of three-dimensional magnetotelluric response: Geophysics. 4(), 182-197.

Tripp. A. C.. and Hohmann. G. W.• 1984. Block diagonalization of the electromagnetic impedance matrix of a symmetric buried body using group theory: IEEE Trans. Geosci. Remote Sensing. 22. 62-69.

Wannamaker. P. E .. Hohmann, G. W., and SanFilipo. W. A.. 1984a. Electromagnetic modeling of three-dimensional bodies in layered earths using integral equations: Geophysics, 49, 60-74.

Wannamaker. P. E.. Hohmann, G. W., and Ward. S. H., 1984h, Magnetotelluric responses of three-dimensional bodies in layered earths: Geophysics. 49, 1517-1533.

Zonge, K. L., and Hughes. L J .• 1988. Controlled source audio­frequency rnagnetotellurics , in Nabighian, M. N.. Ed. Electro­magnetic methods in applied geophysics. 2. Practice: Soc. Explor. Geophys.

Zonge, K. L.. Ostrander. A. Goo and Emer, D. F.• 1986. Controlled source audio-frequency magnetotelluric measurements, in Vo­zotf. K.. Ed .. Magnetotelluric methods: Soc. Explor. Geophys .. 749-763.