magnetovariational sounding: new possibilities

701

Izvestiya, Physics of the Solid Earth, Vol. 39, No. 9, 2003, pp. 701–727. Translated from Fizika Zemli, No. 9, 2003, pp. 3–30.Original Russian Text Copyright © 2003 by Berdichevsky, Dmitriev, Golubtsova, Mershchikova, Pushkarev.English Translation Copyright © 2003 by

MAIK “Nauka

/Interperiodica” (Russia).

INTRODUCTION

Modern magnetotellurics consists of two interre-lated approaches: (i) the magnetotelluric sounding(MTS) based on the simultaneous measurements of theelectric (telluric) and magnetic fields of the Earth and(ii) the magnetovariational sounding (MVS) restrictedto the measurements of magnetic field variations alone[Rokityansky, 1982; Berdichevsky and Zhdanov, 1984;Vozoff, 1991; Berdichevsky and Dmitriev, 2002].

Main MTS characteristics are the impedance tensor[

Z

], determined from relations between the horizontalcomponents of the electric and magnetic fields,

, (1)

where

Eτ Z[ ] Hτ=

Eτ Eτ Ex Ey,( ) Z[ ] Zxx Zxy

Zyx Zyy

,= =

Hτ Hτ Hx Hy,( ),=

and the apparent resistivities

(2)

calculated from the moduli of the off-diagonal compo-nents of the impedance tensor.

The main characteristic of the MVS and MVP meth-ods is the tipper [

W

] (the Wiese-Parkinson vector),determined from the relations between the verticalcomponent of the magnetic field and its horizontalcomponents

, (3)

where

In the traditional scheme of electromagnetic sound-ing using the magnetotelluric field, the MTS methodplays a leading role (stratification of the medium, geo-electric regionalization, mapping of subsurface topog-

ρxy

Zxy2

ωµ0------------,= ρyx

Zyx2

ωµ0------------,=

Hz W[ ] Hτ=

W[ ] Wzx Wzy.=

Magnetovariational Sounding: New Possibilities

M. N. Berdichevsky, V. I. Dmitriev, N. S. Golubtsova,N. A. Mershchikova, and P. Yu. Pushkarev

Moscow State University, Vorob’evy gory, Moscow, 119899 Russia

Received May 6, 2003

Abstract

—Geoelectric research provides unique information on the stratification of the Earth and its deepstructures. A basic geoelectric method is believed to be magnetotelluric sounding (MTS), underlain by theinversion of frequency responses of the impedance tensor, which are determined from the electric and magneticfields. The MTS data interpretation is complicated due to near-surface inhomogeneities distorting the electricfield. Modern geoelectric methods allow more or less reliable identification of these distortions and their elim-ination. The MTS method is complemented by magnetovariational profiling (MVP) using the tipper (the Wiese-Parkinson vector), which is determined from the magnetic field alone. The MVP role reduces to the recognitionand localization of geoelectric structures disturbing the horizontal homogeneity of the medium studied. How-ever, the theorem of uniqueness proved in this paper for the 2-D magnetovariational inversion problem indicatesthat the tipper not only reflects the geoelectric asymmetry of the medium but also provides constraints on itsstratification, i.e., enables the construction of layered inhomogeneous sections. Thus, local magnetovariationalsounding (MVS) based on the straightforward inversion of frequency responses of the tipper can be appliedtogether with the MTS. The main distinctive MVS feature is the fact that distortions of the magnetic field causedby near-surface inhomogeneities attenuate with decreasing frequency and do not spoil the information on deepstructures. Evidently, local MVS allows one to avoid the difficulties inherent in the MTS due to the near-surfacedistortion of apparent resistivity curves, which are constructed from frequency responses of the impedance ten-sor. We arrive at the conclusion that MVS can substantially improve the reliability of geoelectric reconstruc-tions. In this respect, several problems related to the combination of the MVS and MTS methods arise. In thispaper, we analyze an MVS–MTS complex in which the MVS plays the role of a basic method and the MTS isemployed for checking and refining MVS data. An algorithm based on the principle of successive inversions isproposed for the integrated interpretation of MVS and MTS characteristics. The paper consists of theoreticaland experimental parts. The uniqueness theorem is proven in the theoretical part for the 2-D MVS inversionproblem. The experimental part of the paper describes model experiments on the interpretation of MVS andMTS synthetic data and presents a geoelectric model of the Cascadian subduction zone constrained primarilyby MVS data.

702

IZVESTIYA, PHYSICS OF THE SOLID EARTH

Vol. 39

No. 9

2003

BERDICHEVSKY

et al

.

raphies, tracing of horizontal variations in the electricalconductivity, and identification of conducting layers inthe crust and upper mantle), whereas the MVS methodis only employed for the recognition and localization ofcontrasting structures and for the determination of theirstrikes [Rokityansky, 1982]. This scheme has beenextensively and, in many respects, successfully appliedthroughout the world, providing unique information onthe Earth’s interiors (porosity, graphitization and sul-fidization, fluid and rheological regimes, dehydration,and melting). Its weak point is the distortion introducedby inhomogeneities of the uppermost layers of theEarth into the electric field and thereby into the appar-ent resistivity. As the frequency decreases, these effectsbecome galvanic and involve the entire range of lowfrequencies, distorting the information on the deepelectrical conductivity. The galvanic effects give rise toa static (conformal) shift of the low-frequency branchesof apparent resistivity curves. Presently, many methodsare used for suppressing these effects [Bahr, 1988;Jones, 1988; Groom and Bailey, 1989; Zinger, 1992;Caldwell

et al.,

2002; Berdichevsky and Dmitriev,2002]. The most frequent among them are varioustypes of averaging, filtering, reductions to high- andlow-frequency references, and model corrections.However, all these techniques involve the risk ofoverly gross approximations or even subjective (andsometimes erroneous) decisions resulting in structuralmisidentifications.

This scheme can be substantially improved by usingthe MVS potentialities. The point is that, with decreas-ing frequency, the current induced in the Earth involvesincreasingly deeper layers and the effects of near-sur-face inhomogeneities on its magnetic field weaken.Using data on the magnetic field alone, one can obtainreasonably reliable information on deep geoelectricstructures.

The studies in the field of integrated interpreta-tion of MTS and MVS data are conducted in twodirections.

1. Methods of magnetic-to-electric field transforma-tion are being elaborated. In this way, the impedancerelated to the TE-mode of the magnetotelluric field isdetermined. The “induction” curves of apparent resis-tivity constructed with the use of the TE-impedancehave nearly undistorted low-frequency branches. Theidea of such a transformation was proposed byL.L. Vanyan [Osipova

et al.

, 1982]. The first experi-ments in this direction were conducted in the early1980s [Osipova

et al.

, 1982; Bur’yanov

et al.

, 1983]. Inrecent years, L.L. Vanyan, I.M. Varentsov, N.G. Gol-ubev, and E.Yu. Sokolova have developed algorithmsand computer programs for the 2-D TE-impedancedetermination and carried out the interpretation of theinduction curves of apparent resistivity obtained for thewestern coast of the United States [Vanyan

et al.

, 1997,1998]. These investigations played an important role inthe progress of the EMSLAB experiment, because they

substantiated the presence of the asthenosphere in thecontinental part of the Cascadian subduction zone. Themain drawback of such an approach consists in modelerrors caused by the necessity of a priori specificationof a normal (undistorted) impedance. Therefore, theconstruction of induction curves of apparent resistivityshould primarily be regarded as a means for the visual-ization of MVS data facilitating their analysis (qualita-tive assessment of the stratification of the medium andgeoelectric regionalization).

2. The theory and methods of straightforward MVSdata inversion that is based on the minimization ofTikhonov’s functional containing the tipper misfit arebeing developed. This approach goes back to the mag-netotelluric experiments that were carried out in 1988–1990 in the Kyrgyz Tien Shan by geophysical teams ofthe Institute of High Temperatures, Russian Academyof Sciences [Trapeznikov

et al.

, 1997; Berdichevskyand Dmitriev, 2002]. These measurements were madeon a profile characterized by strong local and regionaldistortions of apparent resistivities that complicate theinterpretation of resulting data. The situation could benormalized only with MVS soundings. Figure 1 plotsthe real tippers

Re

W

zy

and the geoelectric model fittingthese curves. The model crust contains an inhomoge-neous conducting layer (a depth interval of 25–55 km)and vertical conducting zones confined to the knownfaults (the Nikolaev line and the Atbashi-Inylchikfault). The figure also presents the model constructedfrom seismic tomography data. The geoelectricmodel is in good agreement with the observationsand coincides remarkably well with the seismicmodel: low resistivities are reliably correlated withlower velocities. This correlation confirms the valid-ity of geoelectric reconstructions based on MVSdata. The experimental investigations in the TienShan indicate that MVS not only localizes crustalstructures but also provides constraints on the stratifica-tion of the crust. Developing this inference, we proposea new MVS scheme in which the MVS plays a leadingrole, whereas MTS data are utilized for a check and amore detailed specification of the MVS results.

This paper is devoted to the problems related to thesecond direction of research in the MVS theory andmethods. We prove the uniqueness of the 2-D MVSdata inversion, describe model experiments on jointMVS and MTS data interpretation, and present a newmodel of the Cascadian subduction zone constructedaccording to a scheme of successive inversions domi-nated by MVS data.

UNIQUENESS OF THE MVS INVERSION PROBLEM SOLUTION

The MVS inversion problem is meant here as thereconstruction of the electric conductivity

σ

(

x

,

y

,

z

)

from tipper values specified within a wide frequencyrange on sufficiently long profiles or in a sufficientlylarge area. The uniqueness of the solution of the inverse

IZVESTIYA, PHYSICS OF THE SOLID EARTH

Vol. 39

No. 9

2003

MAGNETOVARIATIONAL SOUNDING: NEW POSSIBILITIES 703

problem is a key problem of the MVS method, deter-mining its informativeness.

At first glance, MVS anomalies do not seem to pro-vide any information on the normal structure of themedium

σ

N

(

z

),

because we have

W

zx

=

W

zy

= 0 in a hor-izontally homogeneous model. However, in the case ofa horizontally inhomogeneous medium, MVS can beconsidered as frequency sounding that uses the mag-netic field of a buried local source. Such a source can berepresented by any inhomogeneity

∆σ

(

x

,

y

,

z

)

in whichan excessive electric current spreading in the medium isinduced. Evidently, the distribution of this current, aswell as its magnetic field, depends not only on the inho-mogeneity structure

∆σ

(

x

,

y

,

z

)

but also on the normalstructure of the medium

σ

N

(

z

)

. Thus, the solution of theinverse MVS problem

σ

(

x

,

y

,

z

) =

σ

N

(

z

) +

∆σ

(

x

,

y

,

z

)

exists and we should examine its uniqueness.

We consider the 2-D model shown in Fig. 2. In thismodel, a horizontally layered Earth having the normalconductivity

(4)

contains a 2-D inhomogeneous region

S

having anexcessive conductivity

∆σ

(

y

,

z

) =

σ

(

y

,

z

) –

σ

N

(

z

)

. Theinhomogeneity is elongated along the

ı

axis and itscross section has a maximum size

d

. The functions

σ

N

(

z

)

and

∆σ

(

y

, z) are piecewise-analytical. The regionS is underlain at a depth H by an homogeneous base-ment of conductivity σH = const. The air is an ideal insu-lator. The magnetic permittivity is equal everywhere toits vacuum value µ0. The model is excited by a planeelectromagnetic wave striking vertically the Earth’s

σN z( )σ z( ), 0 z H≤ ≤σH, H z≤

=

–0.4

0.4

0

ObservedModeled

T = 25 s

ReWzy

0

–0.4

–0.2

0.2

0

–0.4

–0.2

0.2

0

–0.4

–0.2

0.2

0 100 200 y, km

T = 100 s

T = 400 s

T = 1600 s

(a)

1020100

1000

5030 15 7

500

50

10

7

0 40 80 120 160 200

10

25

40

55

100

y, km

z, km

AIFNL

0 40 80 120 160 200 y, km4.9 5.1 4.9 4.7 4.5 4.76.0 5.8 6.0 6.2 6.0 5.8 6.0 5.8 6.0 6.2

6.5 6.3 6.5 6.3 6.1

6.06.26.46.66.86.66.4

5.95.75.55.5

8.3

8.3 8.1 7.9

1020

30

50

75

150z, km

(b)

(c)

Fig. 1. Magnetovariational sounding in the Kyrgyz Tien Shan Mountains. (a) Plots of the real tipper constructed along a profilecrossing the Kyrgyz Tien Shan. (b) The resistivity section constructed from MVS data [Trapeznikov et al., 1997]: NL, Nikolaevline; AIF, Atbashi-Inylchik faults. The resistivity values (in Ω m) are given within blocks; the lower-resistivity crustal zone (ρ ≤50 Ωm) is shaded. (c) The velocity section from seismic tomography data [Roecker et al., 1993]. Values of P wave velocities (inkm/s) are given within blocks; the low-velocity crustal zone (νp ≤ 6.2 km/s) is shaded.

704

IZVESTIYA, PHYSICS OF THE SOLID EARTH Vol. 39 No. 9 2003

BERDICHEVSKY et al.

surface. The time dependence of the field is describedby the factor e–iωt.

The uniqueness theorem is formulated for the MVSinverse problem in this model as follows: the piece-wise-analytical distribution of the electrical conduc-tivity

(5)

is uniquely defined by the exact values of the tipper

, (6)

that are given on the Earth’s surface z = 0 at all pointsof the Û axis from –∞ to ∞ for all frequencies rangingfrom 0 to ∞.

The proof of the uniqueness theorem consists oftwo stages. First, we consider the asymptotic behav-ior of the tipper Wzy(y) at great distances from theinhomogeneity S and show that the frequencyresponse of this asymptotics uniquely determines thenormal conductivity distribution σN(z). Further, weprove that, if σN(z) is known, the tipper uniquelydetermines the impedance of the inhomogeneousmedium.

An anomalous magnetic field on the Earth’s sur-face can be represented as the field produced in ahorizontally homogeneous layered medium by

σ M( )σN z( ) M S∉σN z( ) ∆σ y z,( )+ M S∈

=

Wzy y( ) = Hz y z 0=,( )Hy y z 0=,( )------------------------------, ∞– y ∞,< < 0 ω≤ ∞<

excess currents of the density jx that are induced inthe region S:

(7)

where hy(y, M0) and hz(y, M0) are the magnetic fields onthe surface of the horizontally homogeneous mediumproduced by a linear current of unit density flowing at apoint M0(y0, z0) ∈ S. These functions have the form[Dmitriev, 1969; Berdichevsky and Zhdanov, 1984]

(8)

(9)

where the factor eλz relates to the upper half-space z ≤ 0and the function U(λ, z, z0) is the solution of the bound-ary value problem

,

(10)

The anomalous magnetic field components and

can be found from the tipper values known at allpoints of the y axis [Dmitriev and Mershchikova,

HyA

y( )Hy

A y z 0=,( )Hy

N z 0=( )------------------------------ jx M0( )hy y M0,( ) S,d

S

∫= =

Hz y( )Hz

A y z 0=,( )Hy

N z 0=( )------------------------------ jx M0( )hz y M0,( ) S,d

S

∫= =

hy y M0,( ) iωµ0---------- λ y y0–( )eλ zcos

0

∞

∫z 0→lim=

× U λ z 0= z0, ,( )λdλ ,

hz y M0,( ) iωµ0----------– λsin y y0–( )eλ z

0

∞

∫z 0→lim=

× U λ z 0= z0, ,( )λdλ ,

d2U λ z z0, ,( )dz2

------------------------------- η2 λ z,( )U λ z z0, ,( )– δ z z0–( ),–=

z z0 0 H,[ ] ,∈,

η λ z,( ) λ2 iωµ0σN z( )– , Reη 0>=

dU λ z z0, ,( )dz

----------------------------- λU λ z z0, ,( )+ 0 at z 0,= =

dU λ z z0, ,( )dz

----------------------------- ηH λ( )U λ z z0, ,( )– 0 at z H ,= =

ηH λ( ) λ2 iωµ0σH– , ReηH 0.>=

HyA

Hz

z

xy

S

σN(z)

σH

σ(y, z)

d H

Fig. 2. Model of a horizontally layered medium containinga 2-D inhomogeneous region S.



2002]. In order to determine , we solve the integralequation

(11)

and, once is found, we easily obtain

(12)

Returning to (8) and (9), we find the asymptoticbehavior of the functions hy(y, M0) and hz(y, M0) at|y – y0| ∞, starting with the function hy(y, M0). Atlarge values of |y – y0|, harmonics with low spatialfrequencies λ make the major contribution to theexpansion of (8). Expanding U(λ, z = 0, z0) in powersof small λ, we have

substituting this series into (8) and integrating, weobtain

(13)

In a similar way, we obtain

(14)

In order to write the relationship between and

in the form containing the magnetotelluric imped-ance, we introduce the functions

(15)

The function Vy(z) is the solution of problem (10) atλ = 0. Equations enabling the determination of the

HyA

Wzy y( )HyA

y( ) 1π---

HyA

y0( ) y0dy0 y–

-------------------------

∞–

∞

∫+ Wzy y( )–=

HyA

Hz y( ) Wzy y( ) 1 HyA

y( )+[ ] .=

U λ z 0= z0, ,( ) U λ 0= z 0= z0, ,( )=

+ λ U λ z 0= z0, ,( )ddλ

--------------------------------------λ 0=

…;+

hy y M0,( ) iωµ0----------

U λ 0= z 0= z0, ,( )y y0–( )2

----------------------------------------------=

+ O1

y y0–( )4--------------------

.

hz y M0,( ) 2iωµ0---------- 1

y y0–( )3-------------------- U λ z 0= z0, ,( )d

dλ--------------------------------------

λ 0=

=

+ O1

y y0–( )5--------------------

.

HyA

Hz

Vy z( ) U λ 0= z z0, ,( ),=

Vz z( ) U λ z z0, ,( )ddλ

----------------------------λ 0=

.=

function Vz(z) are obtained by differentiating (10) withrespect to λ and setting λ = 0:

(16)

In this notation, we have

(17)

Returning to (7), we determine the asymptoticbehavior of the anomalous magnetic field at |y –y0| ∞:

(18)

where

is the total excess current in the inhomogeneity and yS

is the coordinate of a point in the central part of its crosssection S. Thus, according to (16), we have

(19)

d2Vz z( )dz2

------------------ iωµ0σ z( )Vz z( )+ 0, z 0 H,[ ] ,∈=

Vz z( )∂dz

----------------z +0=

Vy 0( ),–=

Vz z( )∂dz

----------------z H=

iωµ0σH– Vz H( )– 0.=

hy y M0,( ) iωµ0----------

Vy 0( )y y0–( )2

-------------------- O1

y y0–( )4--------------------

,+=

hz y M0,( ) 2iωµ0----------

Vz 0( )y y0–( )3

-------------------- O1

y y0–( )5--------------------

.+=

HyA

y( ) iωµ0----------Vy 0( )

jx M0( )y y0–( )2

-------------------- Sd

S

∫ iωµ0----------

Vy 0( )y yS–( )2

--------------------= =

× jx M0( ) Sd

S

∫ iωµ0----------

Vy 0( )y yS–( )2

--------------------Jx,=

Hz y( ) 2iωµ0----------Vz 0( )

jx M0( )y y0–( )3

-------------------- Sd

S

∫ 2iωµ0----------

Vz 0( )y yS–( )3

--------------------= =

× jx M0( ) Sd

S

∫ 2iωµ0----------

Vz 0( )y yS–( )3

--------------------Jx,=

Jx jx M0( ) Sd

S

∫=

Hz y( )Hy

Ay( )

---------------2

y yS–( )------------------

Vz 0( )Vy 0( )-------------- 2

y yS–( )------------------

Vz 0( )dVz z( )

dz----------------

z 0=

---------------------------–= =

706


BERDICHEVSKY et al.

in a region located sufficiently far from the inhomoge-neity S (|(y – yS)| d).

In order to show that the ratio / can beexpressed through the normal impedance of the Earth,we introduce the function

(20)

According to (16), this function satisfies the Riccatiequation

(21)

with the boundary condition

(22)

We obtained the well-known problem for the imped-ance of a 1-D medium with the conductivity σN(z) [Ber-dichevsky and Dmitriev, 1991, 2002]. In the modelunder consideration, the function Z(z) is evidently thenormal impedance of the Earth ZN(z). SettingZ(z) = ZN(z) and taking (19)–(22) into account, we findthe far-zone asymptotics

(23)

Hz HyA

Z z( ) iωµ0

Vz z( )dVz z( )

dz--------------------------------.=

dZ z( )dz

-------------- σN z( )Z2 z( )– iωµ0=

Z H( )iωµ0–σH

---------------.=

ZN 0( )iωµ0 y yS–( )

2------------------------------ Hz y( )

HyA

y( )---------------

y yS– d

,–=

which coincides with the known expression for the farfield of an infinitely long linear current [Vanyan, 1965].The normal impedance ZN is connected with the ratio of

the components and of the anomalous magneticfield, which can be found, according to (11) and (12),from values of the tipper Wzy specified at all points ofthe Û axis (–∞, ∞). If the tipper Wzy is given all along theÛ axis, we synthesize the anomalous magnetic field

( , ) and calculate the normal impedance from thefar-zone asymptotics. Knowing the anomalous mag-

netic field ( , ) and the normal impedance ZN, weintegrate the second equation of Maxwell (the Faradaylaw) and extend the longitudinal impedance Z|| to theentire Û axis:

(24)

Thus, the values of Z|| are found from the values of Wzy.A one-to-one correspondence exists between the distri-bution of Wzy and Z||. Consequently, we can apply thetheorem of Gusarov [1981], stating that the Z|| inversionhas a unique solution, and extend this result to the Wzy

inversion. The uniqueness theorem for the 2-D MVSinversion reduces to that for the 2-D magnetotelluricinversion of the TE-mode. Both methods, MVS andMTS, have a common mathematical basis. A distribu-tion of electrical conductivity is uniquely determinedby exact values of impedances or tippers specified at all

Hz HyA

Hz HyA

Hz HyA

Z || y( )Ex y( )Hy y( )-------------- =

1

1 HyA

+---------------- ZN iωµ0 Hz y( ) yd

∞–

y

∫–

.=

z

xy

w

ρ'1 h1

h2

ρ"1

ρ3

ρ2

0 100

10–1

10–2

10–3

T s1 2⁄,10−1 100 101 102

T s1 2⁄,10−1 100 101 102

102

101

103

104

150 km

100 km

50 km

25 km

150 km

100 km

50 km

25 km

ρ||, Ω m|ReWzy|

Fig. 3. Model illustrating MVS sensitivity to variations in the normal structure: = 100 Ω m; = 10 Ω m; ρ2 = 10000 Ω m; ρ3 = 0;

h1 = 1 km; h2 = 24, 49, 99, and 149 km. The parameter of the curves is H = h1 + h2.

ρ1' ρ1"



points of the Earth’s surface for the entire range of fre-quencies.

A similar approach reducing the uniqueness theo-rem for the tipper to that for the impedance tensor canprove useful for the analysis of the 3-D MVS problem.

Finally, we consider a model illustrating the sensi-tivity of the tipper to variations in the normal section(Fig. 3). This three-layer model (ρ1 = 100 Ω m, ρ2 =10 000 Ω m, and ρ3 = 0; h1 = 1 km and h2 = 24, 49, 99,and 149 km) contains a near-surface rectangular inclu-sion 16 km in width with the resistivity = 10 Ω m.Observations are conducted at a point O at a distance of1 km from the inclusion edge. The frequency responsesof the tipper clearly reflect the variations in the depth(H = h1 + h2) to the conducting basement. It is notewor-thy that the curves of the real tipper |ReWzy| and longi-tudinal apparent resistivity ρ|| have similar bell shapes.However, the curves |ReWzy| are somewhat less sensi-tive to variations in H than the curves ρ||.

MODEL EXPERIMENTS ON THE 2-D INTEGRATED INTERPRETATION OF MVS AND

MTS DATA

The inverse problem of electromagnetic soundingusing the magnetotelluric field consists in the determi-nation of the Earth’s electrical conductivity from thedependence of components of the tipper and impedancetensor on the position of the observation point and thefrequency of field variations. This problem is unstableand therefore ill-posed [Dmitriev, 1987; Berdichevskyand Dmitriev, 1991, 2002]. An arbitrarily small error infield characteristics can give rise to an arbitrarily largeerror in the conductivity distribution. The solution ofsuch a problem is meaningful if, using a priori informa-tion on the structure of the medium studied, one limitsthe region of parameters to be found, and an approxi-mate solution of the inverse problem is sought within acompact set of plausible models forming an interpreta-tion model.

A geoelectric interpretation model should reflectcurrent notions and hypotheses on the sedimentarycover, crust, and upper mantle. Depending on theamount of a priori information, the goal of research,and the field characteristics used, the model can eithersmooth or emphasize geoelectric contrasts and incorpo-rate inhomogeneous layers and local inclusions ofhigher or lower electrical conductivity. An approximatesolution of an inverse problem constrained by the inter-pretation model is chosen using criteria ensuring theagreement of the solution with the available a prioriinformation and pertinent characteristics of the field(the criteria of the a priori information significance andmodel misfit). This approach was implemented in theregularized optimization method [Dmitriev, 1987; Ber-dichevsky and Dmitriev, 1991, 2002]. The number ofsuch criteria is determined by the number of character-

ρ1'

istics specifying the field (real, imaginary, amplitude, andphase characteristics). If the inversion uses a few charac-teristics of the field, the problem is multicriterion.

The 2-D integrated interpretation of MVS and MTSdata belongs to the class of multicriterion problems.The determination of the electrical conductivity of theEarth requires data on the TE-mode with the character-istics ReWzy, ImWzy, ρ||, and ϕ|| (real and imaginary tip-pers, longitudinal apparent resistivities, and phases oflongitudinal impedances) and on the TM-mode with thecharacteristics ρ⊥ and ϕ⊥ (transverse apparent resistivi-ties and phases of transverse impedances). Theseparameters differ in sensitivity to target geoelectricstructures and in stability with respect to near-surfacedistortions [Berdichevsky and Dmitriev, 2002]. TheTE-mode is more sensitive to deep conducting struc-tures and less sensitive to the resistance of the lithos-phere, whereas the TM-mode is less sensitive to deepconducting structures and more sensitive to the resis-tance of the lithosphere. In addition, note that apparentresistivities in the entire range of low frequencies canbe subjected to strong static distortions due to local 3-D near-surface inhomogeneities (geoelectric noise),whereas low-frequency tippers and impedance phasesare free from these distortions. An algorithm of the 2-Dbimodal inversion should implement a procedure inwhich the field characteristics in use support and com-plement each other: gaps arising in the inversion of onecharacteristic of the field should be filled through theinversion of another. In inverting various field charac-teristics, one should give priority to the most reliableelements of the model and suppress the least reliableones.

The following two approaches are possible in solv-ing multicriterion inverse problems: (i) parallel (joint)inversion of all field characteristics used and (ii) suc-cessive (partial) inversions of each of the field charac-teristics.

The parallel inversion summarizes all inversion cri-teria related to various field characteristics and reducesin a 2-D variant to the minimization of the Tikhonov’sfunctional

(25)

where the following notation is used: p, vector of thesought-for parameters; Fm, field characteristic in use; y,coordinate of the observation point; ω, frequency; Im,operator determining Fm from the known distribution ofthe conductivity σ; γm, significance coefficient of thecriterion of the model misfit (the deviation Im from Fm);Ω, solution selection criterion (stabilizer) adjusting thesolution to the a priori information; α, regularizationparameter (the significance coefficient of the a prioriinformation); M, number of the field characteristicsused.

γm Fm y ω,( ) Im σ( )– 2

m 1=

M

∑ αΩ σ( )+

pinf ,

708


BERDICHEVSKY et al.

At first glance, the parallel inversion seems to be themost effective because it incorporates simultaneouslyall specific features of the multicriterion problem andsignificantly simplifies the work of the geophysicist.However, this approach is open to criticism.

If various characteristics Fm have the same sensi-tivity to all parameters p(p1, p2, …, ps ) of the geo-electric structure, their parallel inversion is not veryadvantageous because one, the most reliably deter-mined, characteristic is sufficient for a comprehen-sive inversion.

The use of several characteristics of the field makesthe inversion more informative if they differ signifi-cantly in their sensitivity to various parameters of thegeoelectric structure. However, in this case their jointinversion can become conflicting, because they pro-vide different constraints on the geoelectric structureand are related to different criteria of model misfitsand solution selection. It is possible that in some casesa fortunate selection of weights allows one to con-struct a self-consistent model with a small overall mis-fit. However, the reasonable selection of such weightsis itself a complex problem that often cannot be solvedas yet. Apparently, a succession of partial inversions isthe best approach to the solution of a multicriterioninverse problem.

Let a field characteristic Fm be the most sensitive tothe vector of parameters p(m). Then, the partial mthinversion of the multicriterion 2-D problem consists inthe minimization of the following Tikhonov’s func-tional on the set of the parameters p(m), with the param-eters p – p(m) being fixed:

(26)Fm y ω,( ) Im σ( )– 2 αΩ σ( )+ p m( )inf .

The successive application of the field characteris-tics Fm (m = 1, 2, …, M) reduces the solution of the mul-ticriterion problem to a succession of partial inversions.Each partial inversion is intended for the solution of aspecific problem and can be restricted to specific struc-tures.

A decrease in the number of parameters minimiz-ing the Tikhonov’s functional significantly enhancesthe stability of the problem. Partial inversions com-prehensively incorporate specific features of the fieldcharacteristics used, their informativeness, and theirconfidence intervals. They admit informationexchange between various field characteristics, enablea convenient interactive dialog, and are easily tested.We believe that this direction of research is mostpromising for further development of methodsdesigned for the integrated interpretation of MVS andMTS data.

The method of partial inversions is corroborated byresults of studies carried out in various geological prov-inces [Trapeznikov et al., 1997; Berdichevsky et al.,1998, 1999; Pous et al., 2001; Vanyan et al., 2002].Below, we describe model experiments performed bythis method.

Figure 4 presents a 2-D model schematically illus-trating geoelectric structure of the Kyrgyz Tien Shan[Trapeznikov et al., 1997]. This model, referred tobelow as the TS model, includes the following ele-ments: (1) an inhomogeneous sedimentary cover whoseresistivity ranges from 10 to 100 Ω m; (2) an inhomo-geneous upper crust with a resistivity of 105 Ω m in thenorth and 104 Ω m in the south; (3) an inhomogeneousconducting layer in the lower crust (35–50 km) resistiv-ity in which increases monotonically from 10 Ω m inthe south to 300 Ω m in the north; (4) homogeneousconducting zones A, B, and C branching from thecrustal conducting layer (they are regarded as fault

200150100500–50

1

10

3550

150

500

km

km

z

xy

10 100 10 30 100

104105

300 10030

50 30

10

15 10

10

103

10

SN

Crustal conductorρ ≤ 100 Ω m

Fig. 4. The 2-D TS model. The zone of lower crustal resistivities is shaded, and resistivity values (in Ω m) are shown withinblocks.



zones); and (5) a homogeneous, poorly conductingmantle (103 Ω m) underlain by a conducting mantle (10Ω m) at a depth of 150 km. The model is excited by avertically incident plane wave.

The forward problem was solved with the use of aprogram implementing the finite element method[Wannamaker et al., 1987]. Gaussian white noise wasintroduced into the resulting field characteristics; thenoise had standard deviations of 5% for longitudinaland transverse apparent resistivities ρ|| and ρ⊥ , 2.5° forphases of longitudinal and transverse impedances ϕ||

and ϕ⊥ , and 5% for real and imaginary parts of the tip-per ReWzy and ImWzy. To imitate the static shift due tosmall near-surface (3-D) inhomogeneities, the apparentresistivity curves were multiplied by random real num-bers uniformly distributed in the interval from 0.5 to 2.

It is of interest to estimate the frequency intervalwithin which the tipper is unaffected by the upper layerinhomogeneities. Figures 5 and 6 present the frequencyresponses of ReWzy and ImWzy calculated for the TSmodel and for the same model with a homogeneousupper layer of a resistivity of 10 Ω m. Except for a fewpoints (Û = 45 km for ReWzy and Û = –45, 55, and100 km for ImWzy), the ReWzy and ImWzy curves cal-culated for the TS models with inhomogeneous andhomogeneous upper layers coincide at periods ofabout 100 s and even less. These are periods of thelow-frequency interval, in which the MVS near-sur-face effects attenuate and the effects of crustal inho-mogeneities become appreciable.

The integrated interpretation of the synthetic char-acteristics of the field obtained from the TS model wasperformed by the method of partial inversions.

0.2

–0.1

–0.4

0.1y = –45 km

0.2

–0.1

–0.4

y = 0 km

0.2

–0.1

–0.4

y = 20 km

0.1

–0.2

–0.5

y = 55 km

0.1

–0.2

–0.5100 101

y = 90 km

102

T s,

–0.2

–0.5

y = 100 km

0.1

–0.2

–0.5

y = 120 km

0.1

–0.2

–0.5

y = 145 km

0.1

–0.2

–0.5

y = 165 km

0.1

–0.2

–0.5100 101

y = 190 km

102

T s,

ReW

zy

Fig. 5. Frequency responses of ReWzy. Solid and broken lines are the TS models with inhomogeneous and homogeneous upper lay-ers, respectively.

1/2 1/2

710


BERDICHEVSKY et al.

The construction of the interpretation model isthe most important stage of an interpretation [Dmit-riev, 1987; Berdichevsky and Dmitriev, 1991, 2002].The interpretation model should meet the followingtwo requirements: it should be informative (i.e.,reflect the main features of the study section, includ-ing target layers and structures) and it should besimple (i.e., be determined by a small number ofparameters ensuring the stability of the inverseproblem).

It is evident that these requirements are antago-nistic: a more informative model is more complex.Therefore, an optimal model, both simple and infor-mative enough, should be chosen. This is a key pointof the interpretation, determining both the strategyand even, to an extent, the solution of the inverseproblem. The choice of the interpretation model is amatter of the intuition of a researcher and his or herexperience, understanding of the real geological sit-

uation, adherence to traditions, and ability to departfrom traditional approaches. Although the choice ofthe optimal model is subjective, it is restrained by apriori information, qualitative estimates, and reason-able hypotheses on the structure of the medium stud-ied.

Constructing the interpretation model for inversionsof the synthetic TS characteristics of the field, weassumed that the following a priori information on thestudy medium was available: (1) the sedimentary coveris inhomogeneous, with a tentative thickness of 1 km;(2) the consolidated crust is inhomogeneous and cancontain local conducting zones, its resistivity can expe-rience regional variations, and an inhomogeneous con-ducting layer corresponding to the seismic waveguidepossibly exists in its lower part at depths of 35 to 50 km;and (3) the upper mantle consists of homogeneous lay-ers, and its resistivity at depths below 200 km canamount to a few tens of ohm meter units.

ImW

zy

y = –45 km

y = 0 km

y = 20 km

y = 55 km

100 101

y = 90 km

102

0.2

–0.1

–0.4100 101

y = 190 km

102

0.2

–0.1

–0.4

y = 165 km

0.2

–0.1

–0.4

y = 145 km

0.2

–0.1

–0.4

y = 120 km

0.2

–0.1

–0.4

y = 100 km

T s,

Fig. 6. Frequency responses of ImWzy. Solid and broken lines are the TS models with inhomogeneous and homogeneous upper lay-ers, respectively.

1/2



To additionally refine this information, we invertedthe tippers using a smoothing program capable of iden-tifying and localizing crustal conductors. We appliedthe REBOCC program [Siripunvaraporn and Egbert,2000] using, as an initial approximation, a homoge-neous half-space with a resistivity of 100 Ω m. Figure 7presents the TP-1 model, resulting from the inversion of

ReWzy and ImWzy . This model yields clear evidence ofthree local crustal conducting zones A, B, and C (ρ < 30Ω m) but fails to distinctly resolve layers in the crustand upper mantle.

The a priori information complemented with data onlocal crustal conductors provides a reasonable basis for

–180–60

–160

–140

–120

–100

–60

–80

–40

–20

0z, km

–40 –20 0 20 40 60 80 100 120 140 160 180 200 220y, km

A

BC 100030

100

100

10030

3010

0

300 300

300

300

300

100

3000

1000

1000

3000

3000 1000

10000

3000

1000

300

100

30

10

3

1

ρ, Ω m

Fig. 7. The TP-1 model: inversion of ReWzy and ImWzy using the REBOCC program; A, B, and C are conducting zones in the crust(cf. Fig. 4).

1

200150100500–50

2

5

10

20

50

100

200

500

20 Ω m 100 Ω m 1000 Ω m 5000 Ω mkm

y, km

A

B

C

SN

Fig. 8. The interpretation block model for successive partial inversions; starting values of resistivities (in Ω m) are shown withinblocks.

712


BERDICHEVSKY et al.

the construction of a block interpretation model. Thismodel, presented in Fig. 8, consists of 70 blocks of afixed geometry. The concentration of blocks dependson the position and size of tentative structures and ishighest within the sedimentary cover, local crustal con-ductors, and crustal conducting layer. In minimizingTikhonov’s functional, resistivities of the blocks arevaried. The starting values of resistivities are shown inthe model presented in Fig. 8.

Partial inversions of the synthetic characteristics ofthe field were performed in the class of block structureswith the use of the II2DC program [Varentsov, 2002] inthe following succession: (1) ReWzy and ImWzy inver-sion (2) ϕ|| inversion (3) ρ⊥ and ϕ⊥ inversion.All the inversions were conducted automatically.Below, we consider each inversion separately.

1. Inversion of ReWzy and ImWzy. The startingmodel is shown in Fig. 8. Due to the absence of near-surface distortions, the tipper inversion can providereliable constraints on main structures of the mediumstudied. As a result, we obtained the TP-2 model,shown in Fig. 9, which agrees well with the TS model.The divergence between the tippers calculated fromboth models is generally no more than 5–7% in theperiod range from 1 to 10 000 s (Fig. 10). Using theMVS data alone, we successfully reconstructed (in theautomatic regime, without any corrections!) the mostsignificant elements of the study section, including theinhomogeneous sedimentary cover; the local crustalconductors A, B, and C; and the inhomogeneous crustalconducting layer whose resistivity varies from 234 Ω min the north to 16 Ω m in the south (from 300 to 10 Ω m inthe initial model). Also resolved was the contrast

between the nonconductive and conductive mantle(1667 Ω m/109 Ω m in the TP-2 model and1000 Ω m/10 Ω m in the initial TS model). We see thatthe MVS characteristics of the field measured on a200-km profile allowed us not only to detect the localconducting zones but also to determine the stratificationof the section (with an accuracy sufficient for obtaininggross petrophysical estimates).

2. Inversion of j||. At this stage, without goingbeyond the TE-mode, we control the tipper inversionand gain additional constraints on the stratification ofthe section. A difficulty consists in the fact that thecurves of the longitudinal apparent resistivity ρ|| are dis-torted by near-surface 3-D inhomogeneities that creategeoelectric noise and need to be allowed for. Weavoided this difficulty by restricting ourselves to theinversion of the undistorted curves ϕ||. If ρ|| and ϕ|| areinterrelated through dispersion relations, the disregardof the curves ρ|| does not lead to a loss of information.We interpreted the curves ϕ|| using as a starting modelthe TP-2 model, obtained from the tipper inversion. Weremind the reader that phase inversion gives a resistiv-ity distribution accurate to an unknown scalar factor (asis evident from the analysis of 1-D models). To elimi-nate this uncertainty, we fixed the resistivities of thesedimentary cover and the upper nonconductive crust.Figure 11 presents the TE model, obtained from thephase inversion. The divergences between the phasescalculated from the TE and initial TS models do notexceed 2.5° (Fig. 12). Comparing the TE and TP-2models, we see that the phase inversion agrees reason-ably well with the tipper inversion. Two points are ofparticular interest here: (1) the boundary resistivities of

–50 y, kmSN

1

2

5

10

20

50

100

200

km500


0 50 100 150 200

10 26 156 139 138 138 139 135 11 10 10 9 10 9 32 36 29 115 117 185

5000

2595

1667

109

20

234 136 67 72 33 25 24 13 43 16

9

5181 420 51

28 40

13 6

13 1892 200

>1000

162922

15

12

78

15

10

>10

00

>10

00

5000

AB

C

90

Fig. 9. The TP-2 model: inversion of ReWzy and ImWzy using the II2DC program; resistivity values (in Ω m) are shown withinblocks, and the region of lower crustal resistivities is shaded (cf. Fig. 4).



the inhomogeneous crustal layer (343 and 10 Ω m)became closer to their true values (300 and 10 Ω m),and (2) the contrast between the nonconductive andconductive mantle became sharper (3801 Ω m/15 Ω min the TE model versus 1000 Ω m/10 Ω m in the initialTS model). Thus, phase inversion significantlyimproves the accuracy of the section stratification.

3. Inversion of r^ and j^. This inversion isintended for estimating the resistivity of the upper,highly resistive crust ρupper. The TE model, obtainedfrom the ϕ|| inversion, was used as a starting model.Focusing the inversion of ρ⊥ and ϕ⊥ on the upper crust,we fixed all resistivities except for ρupper . As a result, theinversion of ρ⊥ and ϕ⊥ yielded the TM model, shown inFig. 13. The divergences between the apparent resistiv-ities ρ⊥ calculated from the TM and initial TS models

are seen from Fig. 14. Both models yield similarregional variations of ρ⊥ with a local scatter associatedwith the geoelectric noise. The phase misfits of the TMmodel do not exceed 2.5°. The TM model reveals theasymmetry of the highly resistive upper crust: from283 000 Ω m in the north to 13 000 Ω m in the south (inthe initial TS model, from 105 Ω m in the north to104 Ω m in the south).

The TM model is the final model obtained from thesuccessively applied automatic partial inversions. Itsagreement with the initial TS model is evident (cf.Figs. 13 and 4). All of the major TS structures are wellresolved in the TM model. Misfits between these mod-els do not exceed 5–7% in tippers and 2.5° in phases.

For comparison, Fig. 15 presents the PI model,obtained by the parallel (joint) inversion of all charac-

ReWzy

y, km

T = 1 s

T = 25 s

0.4

0.2

0

–0.2

–0.4–0.6

0.4

0.2

0

–0.2

–0.4

–50 0 50 100 150 200 250

0.1

0

–0.1

–0.2

–0.3

–0.4

0

–0.02

–0.04

–0.06

–0.08

T = 600 s

T = 10000 s

ImWzy

T = 1 s0.4

0.2

0

–0.2

–0.4

0.2

0.1

0

–0.1

–0.2

–0.3

T = 25 s

y, km–50 0 50 100 150 200 250

0

0.02

0.04

0.06

–0.02

T = 10000 s

0.1

0

–0.1

T = 600 s

0.08

TS model TP-2 model

Fig. 10. Comparison of frequency responses of ReWzy and ImWzy calculated from the TS and TP-2 models.

714


BERDICHEVSKY et al.

teristics of the field (ReWzy , ImWzy, ϕ||, ρ⊥ , and ϕ⊥ ) usedfor the construction of the TM model. The startingmodel used here was the same as in the tipper inversion(Fig. 8). The following features characterize the PImodel: (1) resistivity contrasts in the sedimentarycover are significantly smoothed; (2) the resistivitycontrast in the upper, highly resistive crust is also sig-nificantly smoothed (86 000 Ω m/44600 Ω m versus283602 Ω m/13278 Ω m in the TM model and100000 Ω m/10000 Ω m in the initial TS model);(3) the northern conducting zone A is well resolved, thesouthern conducting zone C is somewhat less expressed,and the central conducting zone B is strongly distorted(this zone, connecting the sedimentary cover with thecrustal conducting layer, is only represented by fourconducting blocks); (4) the contrast between the twoboundary resistivities in the crustal conducting layer ismuch lower (120 Ω m/29 Ω m versus 343 Ω m/10 Ω min the TM model and 300 Ω m/10 Ω m in the initial TSmodel); and (5) the monotonic decrease of resistivitiesassociated with the nonconductive/conductive mantletransition is disturbed (a poorly conducting layer with aresistivity of 7423 Ω m is present in the conductingmantle). We see that the parallel inversion of all charac-teristics of the field worsened significantly the interpre-tation result.

Of course, the parallel inversion of all characteris-tics of the field is the simplest approach to a multicrite-rion problem, and apparently this is the reason why it ispopular among geophysicists fascinated by the possi-bility of automatic inversions eliminating the necessityof comprehensive analysis of information. The transi-tion to the technique of successive partial inversions

undoubtedly complicates the work of the geophysi-cist, and this is a possible reason for the objectionsraised in discussions of inverse multicriterion prob-lems of deep geoelectric studies. However, ourexperiments on the integrated interpretation of MVSand MTS data indicate that the game, albeit more dif-ficult, is worth the candle.

MVS STUDY OF THE CASCADIAN SUBDUCTION ZONE (EMSLAB EXPERIMENT)

The above scheme of partial inversions of MVS andMTS data was applied to the construction of the geo-electric model for the Cascadian subduction zone[Vanyan et al., 2002]. We used data obtained in 1986–1988 by geophysicists from the United States, Canada,and Mexico on the Pacific North American coast withinthe framework of the experiment “ElectromagneticStudy of the Lithosphere and Asthenosphere beneaththe Juan de Fuca Plate” (EMSLAB) [Wannamakeret al., 1989a].

Figure 16 presents a predictive petrological and geo-thermal model of the Cascadian subduction zone alongan E–W profile generalizing modern ideas and hypoth-eses on the structure of the region and its fluid regime[Romanyuk et al., 2001]. The subducting Juan de Fucaplate originates at an offshore spreading ridge (about500 km from the coast). In the eastward direction, theprofile crosses (1) an abyssal basin with a sedimentarycover 1–2 km thick and a pillow lava layer 1.5–2 kmthick; (2) the Coast Range, composed of volcanic-sedi-mentary rocks; (3) the Willamette River valley, filledwith a thick sequence of sediments and basaltic intru-

–50 y, kmSN

1

2

5

10

20

50

100

200

km500


0 50 100 150 200

10 26 156 139 138 138 139 135 11 10 10 9 10 9 32 36 29 115 117 185

5000

5716

3801

15

10

343 62 84 98 36 27 14 8 161 10

25

4751 301 91

8 157

8 12

18 17157

>1000

52215

7

11

75

13

11

>10

00

>10

00

5000

A B C

90

Fig. 11. The TE model: inversion of ϕ|| using the II2DC program; resistivity values (in Ω m) are shown within blocks, and the regionof lower crustal resistivities is shaded (cf. Fig. 4).



sions; (4) the Western (older) and Eastern (younger)Cascade ranges, consisting of volcanic and volcanic-sedimentary rocks typical of a recent active volcanicarc; and (5) the Deschutes Plateau, covered withlavas.

The abyssal basin is characterized by a typical oce-anic section with the asthenosphere at a depth of about40 km (the 900°C isotherm). The continental crustabove the subducting slab has lower temperatures. Asubvertical zone of higher temperatures reaching themelting point of wet peridotite (~900°C) has beenlocalized beneath the High Cascades. The release offluids from the upper part of the slab appears to be dueto a few mechanisms. First, at depths to about 30 km,free water is released from micropores and microfrac-tures under the action of the increasing lithostatic pres-sure. Dehydration of minerals such as talc, serpentine,and chlorite starts at depths of 30 to 50 km, where thetemperature exceeds 400°C. Finally the basalt–eclogitetransition can start at depths greater than 75 km, and

exsolution of amphibolites can take place at depths ofmore than 90 km. All these processes are accompaniedby the release of fluids. Supposedly, fluids released atsmall depths migrate through the contact zone betweenthe oceanic and continental plates. At greater depths,fluids can be absorbed by mantle peridotites (serpenti-nization) and, given high temperatures, disturb theequilibrium state of material and give rise to “wet”melting. The melts migrate upward toward the Earth’ssurface, producing a volcanic arc.

Two 2-D geoelectric models of the Cascadiansubduction zone constructed along the Lincoln line(an E–W profile in the middle part of Oregon) havebeen discussed in literature: EMSLAB-I [Wannama-ker et al., 1989b] and EMSLAB-II [Varentsov et al.,1996].

The EMSLAB-I model, shown in Fig. 17a, was con-structed by a trial-and-error method with a strong prior-ity given to the TM-mode; the latter, in the opinion ofthe authors of this model, is least sensitive to deviations

TS model TE model

ϕ||, deg

–10

–20

–80

–65

–20

–30

–30

–30

–40

–40

–40

–40

–40

–40

–50

–50

–50

–50

–50

–50

–50

–60

–60

–60

–60

–60

–60

–70

–70

–70

–70

–70

–80

–70–50 0 50 100 150 200

y, km

T = 0.25 s

–50 0 50 100 150 200

T =1 s

T = 6 s

T = 25 s

y, km

T = 10000 s

T = 2500 s

T = 600 s

T = 100 s

ϕ||, deg

Fig. 12. Comparison of frequency responses of ϕ|| calculated from the TS and TE models.

716


BERDICHEVSKY et al.

ρ⊥ , Ω m

y, km

100

101

102

103

104

102

101

100

103

102

101

100

103

102

101

100

104

103

102

101

100

104

103

102

101

100

104

103

102

101

100

10–1

103

102

101

100

10–1

–50 0 50 100 150 200–50 0 50 100 150 200

T = 0.25 s

T = 1 s

T = 6 s

T = 25 s

T = 100 s

T = 600 s

T = 2500 s

T = 10000 s

ρ⊥ , Ω m

y, km

TS model TM model

157

–50 y, kmSN

1

2

5

10

20

50

100

200

km500


0 50 100 150 200

10 26 156 139 138 138 139 135 11 10 10 9 10 9 32 36 29 115 117 185

13 000

5716

3801

15

10

343 62 84 98 36 27 14 8 161 10

25

4751 301 91

8 157

8 12

18 17

>1000

52215

7

11

75

13

11

>10

00

>10

00

283 000

A B C

90

Fig. 13. The TM model: inversion of ρ⊥ and ϕ⊥ using the II2DC program; resistivity values (in Ω m) are shown within blocks, andthe region of lower crustal resistivities is shaded (cf. Fig. 4).

Fig. 14. Comparison of frequency responses of ρ⊥ calculated from the TS and TM models.



–50 y, kmSN

1

2

5

10

20

50

100

200

km500


0 50 100 150 200

18 30 75 78 83 83 82 78 32 29 21 21 35 37 43 45 45 66 71 90

44 600

904

160

7423

9

120 481 77 120 99 24 33 37 77 29

65

156610 41

121

23 11108

>1000 7

270

5

150

601

246

86 000

A B C

90

56

39223

199 97 >1000 >1000

>1000

>1000 >1000

Fig. 15. The PI model: parallel inversion of ReWzy, ImWzy, ϕ||, ρ⊥ , and ϕ⊥ using the II2DC program; resistivity values (in Ω m) areshown within blocks, and the region of lower crustal resistivities is shaded (cf. Fig. 4).

Fig. 16. Predictive geothermal and petrological CASCADIA model, constructed along an E–W profile across central Oregon[Romanyuk et al., 2001].

50

100

150

Dep

th, k

m

Abyssalbasin 100 200 400 Distance, km

250°ë

500°ë

750°ë

900°ë

1250°ë750°ë

H2O

H2O

750°ë

900°ë

500°ë900°ë (dry lherzolite solidus)

Harzburgite replaced with depthby spinel-garnet lherzolite

Asthenosphere (lherzolite with ~2% melt)

Neogene sediments of a continental basin

Loose/compacted oceanic sediments

Post-Eocene accreted sediments

Pre-Eocene accreted oceanic sediments

Oceanic basalts/gabbros

Mafic eclogites

Pre-Eocene continental crust, mostly composed of accreted terranes

Late Eocene–Mid-Miocene volcanic–sedimentary complex of the

Late Eocene–Mid-Miocene basalts/gabbroids intruded

Neogene andesitic basalts

Recent amphibolite/granulite metamorphism

Melting of wet/dry peridotites

Subducting sediments and fragments of the oceanic and continental crust

Prehnite–pumpellyite/blueschist metamorphism in the contact zone

Melting of wet acidic rocks/eclogitization of basic rocks

Serpentinization of mantle peridotites in the mantle wedge

Volcanoes

and subjected to Tertiary amphibolite/granulite metamorphism

Late Paleocene–Mid-Eocene mafic basalts and gabbroid rocks intrudedinto the accretionary prism or coastal portions of the pre-Tertiary crust

Western Cascades

into the pre-Tertiary crust

250°ë

Coast CoastRange

WillametteValley

WesternCascades

High Cascades

718


BERDICHEVSKY et al.

from the 2-D pattern. The EMSLAB-I model mini-mizes the misfits of the curves ρ⊥ and ϕ⊥ and ignoresthe curves ρ|| and ϕ||. Its main elements are (1) the upperconductive part of the plate, sinking at a low anglebeneath the Coast Range; (2) a subhorizontal conduct-ing layer in the middle continental crust broadening inthe area of the High Cascades; and (3) a well-developedconductive asthenosphere beneath the ocean. The prob-lem of the junction between the slab and the crustal

conductor remains open in this model. The continentalasthenosphere is not present in this model, although theshape of the experimental curves ρ|| and ϕ|| suggests alow resistivity of the upper mantle. The absence of largedivergences between the model values of ReWzy andImWzy, on the one hand, and the experimental data, onthe other hand, is considered by the authors as evidenceof the reliability of the model.

Fig. 17. Geoelectric models of the Cascadian subduction zone: (a) EMSLAB-I [Wannamaker et al., 1989b]; (b) EMSLAB-II [Va-rentsov et al., 1996]. CB, Cascadia basin; NB, Newport basin; CR, Coast Range; WV, Willamette Valley; WC, Western Cascades;HC, High Cascades; DP, Deschutes Plateau.

0

5

25

50

100

150200

300

400km

150 100 50 0 50 100 150 200

250

kmDPHCWCWVCRNBCB

(a)

10–1200 150 100 50 0 50 100 150 km

(b)HCWCWVCRNBCB

100

101

102

kmResistivity, Ω m

<1 <3 <10 <30 <100 <300 <1000 <3000

10–1

100

101

102

km

4 : 1

1 : 1

1 : 4



The EMSLAB-I model is open for criticism. A coldcontinental mantle contradicts current geodynamicideas of the Cascadian subduction zone (compare theEMSLAB-I model with the predictive CASCADIAmodel shown in Fig. 16). Analysis of the EMSLAB-Imodel has shown that the TM-mode is weakly sensitiveto variations in the electrical conductivity of the mantleand that the bimodal inversion alone, using both the TEand TM modes, can be effective in studying theasthenosphere [Berdichevsky et al., 1992; Vanyanet al., 2002].

Experiments on the bimodal interpretation of MTSand MVS data obtained in the Cascadian subductionzone resulted in the construction of the 2-D EMSLAB-IImodel (Fig. 17b). It was constructed with the use of theautomatic inversion program INV2D-FG, optimizingresistivities on 20 blocks of a fixed geometry[Varentsov et al., 1996]. An algorithm of parallelweighted inversion was applied to ϕ⊥ , ReWzy, andImWzy (maximum weight), ϕ|| and ρ⊥ (normal weight),and ρ|| (minimum weight). The EMSLAB-II model hasmuch in common with EMSLAB-I. Both have the same

oceanic asthenosphere, subducting slab, and crustalconducting layer. However, the EMSLAB-II plate joinsthe crustal conductor, and a conducting asthenosphereis present in the EMSLAB-II continental mantle. Thus,the geoelectric data have revealed partial melting in thecontinental mantle. The main drawback of theEMSLAB-II model is its sketchiness due to the limitedpossibilities of the INV2D-FG program.

Presently, the INV2D-FG program has given way tomore efficient software tools designed for the automatic2-D inversion of MVS and MTS data. These are thesmoothing program REBOCC [Siripunvaraporn andEgbert, 2000] and the programs IGF-MT2D[Nowozynski and Pushkarev, 2001] and II2DC [Var-entsov, 2002]. These programs enable the optimizationof models containing 512 and more blocks of a fixedgeometry and provide new possibilities for interpretingthe EMSLAB experimental data [Vanyan et al., 2002].

Three-dimensional model estimates obtained for thePacific coast of North America and analysis of experi-mental field characteristics, induction arrows, and polardiagrams show that the regional structure along the Lin-

–70–50 0 50 100 150 200 y, km

–60

–50

–40

–30

–20

–10

0z,

km

CR1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

WV WC HC DP

1230

869

614

434

307

217

154

109

76.8

54.3

38.4

27.2

19.2

13.6

9.60

6.79

4.80

3.39

2.40

1.70

1.20

0.85

0.60

0.42

0.30

ρ, Ω m

Fig. 18. The 2-D inversion of ReWzy, ImWzy, and ϕ|| performed with the use of the smoothing REBOCC program: CR, Coast Range;WV, Willamette Valley; WC, Western Cascades; HC, High Cascades; DP, Deschutes Plateau.

720


BERDICHEVSKY et al.

coln line is favorable for the 2-D interpretation of MVDand MTS data.

The interpretation consisted of three stages.

At the first stage, the 1-D inversion of short-periodmagnetotelluric (MT) curves (T = 0.01–100 s) was per-formed and an approximate geoelectric section of thecontinental volcanic-sedimentary cover was con-structed to a depth of 3.5 km. This section agrees withthe near-surface portion of the EMSLAB-I model[Wannamaker et al., 1989b].

At the second stage, the REBOCC program wasapplied in experiments with 2-D smoothed inversion.With the complicated conditions of the Cascadian sub-duction zone, the parallel inversion of the TE- and TM-modes yielded whimsical alternation of low- and high-resistivity spots with a poor minimization of the misfit.It is difficult to recognize real structures of the subduc-tion zone in these spots. The most interesting result wasobtained from the partial inversion of ReWzy, ImWzy,and ϕ|| (Fig. 18). The western and eastern conductingzones identified here are separated by a T-shaped regionof higher resistivity that can be associated with the sub-ducting slab. An oceanic asthenosphere whose top canbe fixed at a depth of about 30 km is recognizable in thewestern conducting zone. The eastern conducting zonecoincides with the crust-mantle zone of wet melting inthe predictive CASCADIA model shown in Fig. 16. Itis noteworthy that the upper boundary of the easternconductor closely resembles the topography of the

crustal conducting zone in the EMSLAB-I andEMSLAB-II models shown in Fig. 17.

At the third, final stage, the method of partial inver-sions was applied and a new 2-D geoelectric model ofthe Cascadian subduction zone was constructed [Vanyanet al., 2002]; this model was called EMSLAB-III. Itwas constructed with the use of the II2DC program[Varentsov, 2002], minimizing the model misfit in theclass of media with a fixed geometry of blocks. Thealgorithm of interpretation consisting of partialinversions is illustrated in Fig. 19. The interpreta-tion was conducted in the regime of testing hypoth-eses. We consider three hypotheses on the structureof the Cascadian subduction zone: (1) CASCADIApredictive model, (2) EMSLAB-I model, and(3) EMSLAB-II model.

The interpretation model is shown in Fig. 20a. Theocean floor topography and thicknesses of the seafloor,accretionary prism, and shelf sediments were specifiedfrom bathymetric and sedimentary thickness maps[Connard et al., 1984a, 1984b]. The resistivities of thewater, sediments, and oceanic crust were set at 0.3, 2, and10 000 Ω m, respectively. The depth to the oceanicmantle and its resistivities were taken in accordancewith the CASCADIA, EMSLAB-I, and EMSLAB-IImodels. The slab surface was constructed from seismic[Trehu et al., 1994] and seismic tomography [Weaverand Michaelson, 1985; Rasmussen and Humphries,1988] data. The structure of the volcanic-sedimentarycover was determined from the results of a 1-D inver-sion of short-period MT curves. The crust and mantle of

Fig. 19. Algorithm of interpretation of EMSLAB experimental data in the class of block models.

InitialSTARTmodel

Inversion ofReWzy and TP

model

TEmodel

StartingTP

model

StartingTE

model

Inversion of

ϕ||

TMmodel

Inversion of

ρ⊥ and ϕ⊥

Synthesis of theTP, TE, and TM

models

EMSLAB-IIImodel

ImWzy



the continent were divided into homogeneous blocks.The division density and block geometry were chosenwith regard for the configuration of the eastern con-ducting zone, delineated with the help of the REBOCCprogram, and admitted a free choice of crust and mantlestructures within the framework of the three hypothesesconsidered. A hypothesis best fitting the observed datais chosen automatically in the process of optimizationof resistivities and minimization of misfits. The crustand mantle of the continent have a resistivity of 1000 Ω m

in the starting START model constructed on the basis ofthe interpretation model. Below, we consider the suc-cessive partial inversions.

Inversion of ReWzy and ImWzy. The START modelwas taken as the starting one. The TP model resultingfrom the inversion is shown in Fig. 20b. The tipper mis-fit (the rms deviation of model tippers from observedvalues) in this model is 5–10 times smaller than the tip-per amplitude (the difference between the maximumand minimum tipper values), which is evidence of good

Fig. 20. Models obtained by partial inversions of the EMSLAB experimental data (resistivity values in units of Ω m are shownwithin blocks): (a) interpretation model (resistivity values of the initial START model are shown); (b) TP-2 model (inversion ofReWzy and ImWzy); (c) TE model (inversion of ϕ||); (d) TM model (inversion of ρ⊥ and ϕ⊥ ). CR, Coast Range; WV, WillametteValley; WC, Western Cascades; HC, High Cascades; DP, Deschutes Plateau.

–200 –100 0 50 100 150 200 km

0

33.510

37.5

110

280

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15CR WV WC HC DP

0

0.61.5

25

35

45

70

110

170

280

z, kmz, km

(a)

–200 –100 0 50 100 150 200 km

0

33.510

37.5

110

280

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15CR WV WC HC DP

0

0.61.5

25

35

45

70

110

170

280

z, kmz, km

(b)

104

140

8

1000 1000 1000 1000

1000 1000 1000 1000

1000

1000 1000

10001000

1000

1000

1000 1000 1000 1000

1000100010001000

1000 1000 1000

100010001000

1000

1000

1000 1000

1000

1000

1000

10

15

20

30

40

5

0.3

22

2

4 4

4

10000

10000 10000

10000

90 130 130 1414 40 40 4060 20 11 30

1869

7

30

10010035 35

20

127

6.9

452 47 183 1760

2060 58 72 89

38

23 42

5315490

71

234

147 16 20 71

6920191360

2080 17 38

35221330

5450

337

234

1310

27

22

7.5

8.4

58

53

35

0.79

0.3

1.73.3

5.7

4.7

5

10080

10010 9990

8310

81 120 158 1721 37 34 3471 19 9 33

235.913

11

26

1187849 56

20

4.1

143

18601940

10001000

722


BERDICHEVSKY et al.

agreement of the model with observations. A remark-able feature of the TP model is the conducting conti-nental asthenosphere and the vertical low-resistivityzone that branches from the asthenosphere and crossesthe continental crust in the High Cascade region. Thisfeature distinguishes the TP model from EMSLAB-Iand EMSLAB-II and makes it similar to the predictiveCASCADIA model; in the High Cascades, the latterdelineates a vertical high-temperature zone of wet and drymelting that is evidently characterized by low resistivities.

Inversion of j||. At this stage, we checked the tipperinversion results. To avoid difficulties related to near-

surface distortions of the curves ρ||, we restricted our-selves to the inversion of the curves ϕ||, which satisfythe dispersion relations. The TP model, obtained fromthe tipper inversion, was used as a starting model. Theinversion of longitudinal phases yielded the TE model,shown in Fig. 20c. The phase misfit (the rms deviationof model phases from observed values) in this model is5–10 times smaller than the phase amplitude (the dif-ference between the maximum and minimum phasevalues), indicating good agreement of the model withobservations. As distinct from the TP model, the TEcontinental crust includes a better delineated conducting

Fig. 20. (Contd.)

–200 –100 0 50 100 150 200 km

0

33.510

37.5

110

280

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15CR WV WC HC DP

0

0.61.5

25

35

45

70

110

170

280

z, kmz, km

(c)

–200 –100 0 50 100 150 200 km

0

33.510

37.5

110

280

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15CR WV WC HC DP

0

0.61.5

25

35

45

70

110

170

280

z, kmz, km

(d)

1000030001000300100301031

ρ, Ω m

124

6.5

180 104 6110 2420

116 107 4230 14

16

17 46

19191

14

73

147 12 36 215

61246181280

754 24 42

47271220

4700

299

142

1190

12

22

22

78

22

0.3

1.74.0

2.7

3.4

4

9800

9840 3490

8320

69 54 180 33132 21 19 2259

13 26

6.322

13

23

9313968 108

23

182

63

1650 504 1440 2320

190 90 734 13

38

57 47

1866

11

468

176 119 59 6.7

134091272110

2210 36 47

1835895180

258

244

281

13100

2610

12

8.1

25

25

30

95

69

5

0.3

1.84.4

1.2

3

1660

14 130 1620

2290

82 18 313 70150 20 8.7 4262 35 13 27

1012

8.8

111

4017141 46

20

34

15

30

127

25202030

8.7

4.2

15

5.1

4.8

5.1

502244



layer (ρ = 14–46 Ω m) in a depth interval of 35–45 km,whereas the subvertical conducting zone (ρ = 12–46 Ω m)in a depth interval of 45–110 km, bounded by layerswith resistivities of 147–1260 Ω m to the west and 215–612 Ω m to the east, is localized with a higher contrast.One may state that the TE model is an updated TPmodel.

Inversion of r^ and j^. At this stage, we invertedthe TM-mode, which is less sensitive to conductingzones in the crust and mantle but is more effective inresolving the structure of the junction zone betweenthe slab and crustal conducting layer and providesmore reliable estimates of the resistivity in the upperconsolidated crust. In inverting the TM-mode, the TEmodel, obtained from the inversion of phases ϕ||, wastaken as a starting model. The inversion of transverseapparent resistivities and phases of the transverseimpedance yielded the TM model shown in Fig. 20d.In this model, the misfits of transverse apparentresistivities at most points vary within 6–12%, andthe phase misfits are 7–10 times smaller than thephase amplitude (the difference between the maxi-mum and minimum phase values). The TM modelinherits the main features of the starting TE model(albeit with some deviations). The following impli-cations of the TM model are noteworthy. First, nowell-conducting junction is present between the con-

ducting slab and the crustal conducting layer. Sec-ond, the upper consolidated crust of the continent hasa resistivity of about 2000 Ω m, indicating that it isfractured.

Note that the TM-mode inversion is essentiallydependent on the choice of the starting model. If theSTART model is taken as the initial one, the TM-modeinversion yields a model in which the continentalasthenosphere is absent. This is evidently due to the lowsensitivity of the TM-mode to deep conducting struc-tures. In this case, active is the same mechanism thatgave rise to the model EMSLAB-I, devoid of the con-tinental asthenosphere, in [Wannamaker et al.,1989b].

Synthesis. At this stage, we analyzed the TP, TE,and TM models and constructed the general EMSLAB-III model, smoothing insignificant details and enlargingblocks on the basis of information characterizing theeffect of individual blocks on the MVS and MTS char-acteristics. All changes were made in an interactiveregime, with the calculation of local misfits and withthe correction of boundaries and resistivities. Theresulting general model shown in Fig. 21 provides acoherent geoelectric image of the subduction zone. Theextent of its agreement with observed data is seen fromFig. 22, where the model curves ρ⊥ , ρ||, ϕ⊥ , ϕ||, ReWzy,

V IV III II I 1 2 3 4 5 6 7 8 9 10 11 12 1314 15CR WV WC HC DP

3.03.5

37.5

110

280

z, km–200 –100 0 50 100 150 200 km

0.61.5

40

100

280

z, km

155

25

7.010000

3000

1000

300

100

30

10

3

1

ρ, Ω m

0.3

23

3

4500

30

55

20

1500

15

4

50

50

350 25

250

20

50

88 13 3 7 20

20

7

40

4.5 12 35 14 5 10 200

20

10

2000

2000

20

20 20

20

2060

500 30 500

30

500 (?)

55

Fig. 21. Final EMSLAB-III model (resistivity values in units of Ω m are shown within blocks): I–V, soundings on the ocean floor;1–15, soundings on the continent. CR, Coast Range; WV, Willamette Valley; WC, Western Cascades; HC, High Cascades; DP, Des-chutes Plateau.

724


BERDICHEVSKY et al.

and ImWzy are compared with the observed curves (thestatic distortion in the observed curves ρ|| was removedby a vertical shift of their low-frequency branches). Themodel curves agree well with the observed curves at themajority of points.

In its oceanic part, the EMSLAB-III model is closeto EMSLAB-I and EMSLAB-II and exhibits a thickoceanic asthenosphere in a depth interval of 37.5–110 km. The structure of the continental part ofEMSLAB-III is distinguished by the following signifi-cant elements:

a crustal conducting layer (ρ = 20 Ω m, a depthinterval of 25–40 km) and a conducting asthenosphere(30 Ω m, 100–155 km) are distinctly resolved;

the crustal and asthenospheric conductors are con-nected by a columnlike conducting body (20–30 Ω m)crossing the lithosphere and reaching depths of about7 km in the volcanic zone of the High Cascades;

the subducting slab, in a depth interval of 4–40 kmcontains a thin inclined conductor (20 Ω m) separatedfrom the crustal conducting layer by a higher-resistivityzone (60 Ω m); apparently, the crustal conducting layeris unrelated to slab fluids and has a deep origin.

The reliability of these elements is supported by thefact that the elimination of any of them noticeablyincreases the model misfits.

These features of the continental section make theEMSLAB-III and predictive CASCADIA models sim-ilar. The fluid regime of the subduction zone is clearlyobservable here. The subducting slab entraps fluid-sat-urated low-resistivity rocks of the ocean floor. As theslab moves down, the released free water migratesthrough the shear zone (the contact zone between thesubducting oceanic and stable continental plates). Thedehydration (the release of bound water) developing inthe slab at depths of 30–40 km supplies fluids to themantle and causes the wet melting of asthenosphericmaterial. The low-resistivity melts move upward

101

–0.4

100

Wzy

T s1/2,102

ImWzy

ReWzy

–75

ϕ, d

eg

ϕ||

103

ρ a, Ω

m

–0.2

0

0.2

0.4

0.6

–0.6

–60

–45

–30

–15

0

–90

ϕ⊥

102

101

ρ||

ρ⊥

MTS 1

101 102

ImWzy

ReWzy

ϕ||

ϕ ⊥

ρ||

ρ⊥

MTS 2

101 102

ImWzy

ReWzy

ϕ||

ϕ ⊥

ρ||

ρ⊥

MTS 7

101 102

ImWzy

ReWzy

ϕ||

ϕ ⊥

ρ||

ρ⊥

MTS 8

Fig. 22. Comparison of the observed MTS and MVS curves with the curves calculated from the EMSLAB-III model: (1) observa-tions; (2) EMSLAB-III model.



through the lithosphere and form a volcanic arc. Theheating of the lithosphere activates dehydration in thelower crust, producing the crustal conducting layer.

All this confirms the validity of the predictionsunderlying the CASCADIA model.

CONCLUSION

Theoretical and experimental results indicate thatthe MVS method can play a significant role in thedevelopment of geoelectric studies. Its main advantageconsists in the fact that the magnetic field distortionscaused by near-surface inhomogeneities attenuate withdecreasing frequency and do not spoil the informationon structures of the crust and upper mantle. This makesthe results of electromagnetic studies more reliable andbetter constrained. The development of the MVSmethod should be regarded as a promising problem ofmodern geophysics.

Presently, we are at the very beginning of thisresearch, and many questions remain unanswered.What is the best way to organize qualitative analysis ofMVS data that accounts most adequately for horizontaland vertical variations in the electrical conductivity?What is the resolution of tippers with respect to thesevariations? How rapidly do the magnetovariationalnear-surface effects attenuate at low frequencies? Whatis the sensitivity of tippers to deep structures in the crustand mantle? Which geoelectric conditions are favorablefor the application of the MVS method? What are theconditions at which tippers admit a 2-D approximationof 3-D structures? Such is the list of problems, far frombeing complete, whose solution is crucial for theprogress of MVS.

Answers to many questions can be obtained with thefurther accumulation of experience in studies in variousgeological provinces. Here, we should emphasize that anecessary condition for the MVS application is the

101

–0.4

100

Wzy

T s1/2,102

ImWzy

ReWzy

–75

ϕ, d

eg

ϕ||

103

ρ a, Ω

m

–0.2

0

0.2

0.4

0.6

–0.6

–60

–45

–30

–15

0

–90

ϕ⊥

102

101ρ||

ρ⊥

MTS 9

101 102

ImWzy

ReWzy

ϕ||

ϕ ⊥

ρ||

ρ⊥

MTS 12

101 102

ImWzy

ReWzy

ϕ||

ϕ ⊥

ρ||

ρ⊥

MTS 14

ϕ||

ϕ ⊥

ρ||

ρ⊥

MTS II(ocean)102

101

100

10–1

101 102

1

2

Fig. 22. (Contd.)

726


BERDICHEVSKY et al.

presence of conducting inhomogeneities that play therole of deep local sources of electromagnetic field.

The main goal of this paper was to attract the atten-tion of geophysicists to the potentialities of the MVSmethod and to problems whose solution is indispens-able to the practical realization of these potentialities.

ACKNOWLEDGMENTS

We are grateful to P. Weidelt and U. Schmucker fordiscussions stimulating this work. This work was sup-ported by the Russian Foundation for Basic Research,project nos. 02-05-64079 and 03-05-64167.

REFERENCES

Bahr, K., Interpretation of Magnetotelluric Impedance Ten-sor: Regional Induction and Local Telluric Distortion,J. Geophys., 1988, vol. 62, pp 119–127.Berdichevsky, M.N. and Zhdanov, M.S., Advanced Theory ofDeep Geomagnetic Sounding, Amsterdam: Elsevier, 1984.Berdichevsky, M.N. and Dmitriev, V.I., Magnitotellu-richeskoe zondirovanie gorizontal’no-odnorodnykh sred(Magnetotelluric Sounding of Horizontally HomogeneousMedia), Moscow: Nedra, 1991.Berdichevsky, M.N., Koldaev, D.S., and Yakovlev, A.G.,Magnetotelluric Sounding on an Ocean Coast, Fiz. Zemli,1992, no. 6, pp. 87–96.Berdichevsky, M.N., Dmitriev, V.I., and Pozdnjakova, E.E.,On Two-Dimensional Interpretation of MagnetotelluricSoundings, Geophys. J. Int., 1998, vol. 133, pp. 585–606.Berdichevsky, M.N., Vanyan, L.L., and Koshurnikov, A.V.,Magnetotelluric Sounding in the Baikal Rift Zone, Izvestiya,Phys. Solid Earth, 1999, vol. 35, pp. 793–814.Berdichevsky, M.N. and Dmitriev, V.I., Magnetotellurics inthe Context of the Theory of Ill-Posed Problems, Tulsa: SEG,2002.Bur’yanov, V.B., Gordienko, V.V., Kulik, S.N., and Logvi-nov, I.M., Kompleksnoe geofizicheskoe izuchenie tektonos-fery kontinentov (Multidisciplinary Geophysical Study of theContinental Tectonosphere), Kiev: Naukova Dumka, 1983.Caldwell, T.G., Bibby, H.M., and Brown, C., The Magneto-telluric Phase Tensor—A Method of Distortion Analysis for3D Regional Conductivity Structures, Abstracts of 16thWorkshop on EM Induction in the Earth, Santa Fe, EM12-4.Connard, G., Couch, R., Keeling, K., et al., Abyssal Plainand Continental Net-Objective Sedimentary Thickness,Western North American Continental Margin and AdjacentOcean Floor off Oregon and Washington. Atlas 1 OceanMargin Drilling Program, L.D. Kulm et al., Eds., RegionalAtlas Series: Marine Science International, Sheet 7, 1984a.Connard, G., Couch, R., Pitts, G.S., and Troseth, S., Bathym-etry and Topography, Western North American ContinentalMargin and Adjacent Ocean Floor off Oregon and Washing-ton. Atlas 1 Ocean Margin Drilling Program, L.D. Kulmet al., Eds., Regional Atlas Series: Marine Science Interna-tional. Sheet 1, 1984b.Dmitriev, V.I., Elektromagnitnye polya v neodnorodnykh sre-dakh (Electromagnetic Fields in Inhomogeneous Media),Moscow: MGU, 1969.

Dmitriev, V.I., Inverse Problems in Electrodynamical Pros-pecting, Ill-Posed Problems in the Natural Sciences, Mos-cow: Mir Publishers, 1987, pp. 77–101.Dmitriev, V.I. and Mershchikova, N.A., Synthesis of theMagnetotelluric Field, Fiz. Zemli, 2002, no. 11, pp. 69–75.Groom, R.W. and Bailey, R.C., Decomposition of Magneto-telluric Impedance Tensor in the Presence of Local Three-Dimensional Galvanic Distortion, J. Geophys. Res., 1989,vol. 94, no. B2, pp. 1913–1925.Gusarov, A.L., On the Uniqueness of the MagnetotelluricData Inversion in a Two-Dimensional Medium, Matema-ticheskie modeli v geofizike (Mathematical Models in Geo-physics), Moscow: MGU, 1981, pp. 31–61.Jones, A.G., Static Shift of Magnetotelluric Data and ItsRemoval in a Sedimentary Basin Environment, Geophysics,1988, vol. 53, no. 7, pp. 967–978.Nowozynski, K. and Pushkarev, P.Yu., The Efficiency Anal-ysis of Programs for Two-Dimensional Inversion of Magne-totelluric Data, Izvestiya, Physics of the Solid Earth, 2001,vol. 37, pp. 503–516.Osipova, I.L., Berdichevsky, M.N., Vanyan, L.L., and Bori-sova, V.P., Geoelectric Models of North America, Geomagn.Issled., 1982, no. 29, pp. 117–130.Pous, J., Queralt, P., and Marcuello, A., Magnetotelluric Sig-nature of the Western Cantabrian Mountains, Geophys. Res.Lett., 2001, vol. 28, no. 9, pp. 1795–1798.Rasmussen, J. and Humphries, G., Tomographic Image of theJuan de Fuca Plate beneath Washington and Western OregonUsing Teleseismic P-Wave Travel Times, Geophys. Res.Lett., 1988, no. 15, pp. 1417–1420.Roecker, S.W., Sabitova, T.M., Vinnik, L.P., et al., Three-Dimensional Elastic Wave Velocity Structure of Western andCentral Tien Shan, J. Geophys. Res., 1993, vol. 98, no. B9,pp. 15 579–15 795.Rokityansky, I.I., Geoelectromagnetic Investigations of theEarth’s Crust and Mantle, Berlin: Springer, 1982.Romanyuk, T.V., Mooney, W.D., and Blakely, R.J., A Tec-tonic–Geophysical Model of the Cascadian Subduction Zonein North America, Geotektonika, 2001, no. 3, pp. 88–110.Siripunvaraporn, W. and Egbert, G., An Efficient Data Sub-space Inversion Method for 2-D Magnetotelluric Data, Geo-physics, 2000, vol. 65, pp. 791–803.Trapeznikov, Yu.A., Andreeva, E.V., Batalev, V.Yu., et al.,Magnetotelluric Sounding in the Kyrgyz Tien Shan, Fiz.Zemli, 1997, no. 1, pp. 3–20.Trehu, A.M., Asudeh, I., Brocher, T.M., et al., Crustal Archi-tecture of the Cascadia Forearc, Science, 1994, vol. 265,pp. 237–243.Vanyan, L.L., Osnovy elektromagnitnykh zondirovanii (Fun-damentals of Electromagnetic Sounding), Moscow: Nedra,1965.Vanyan, L.L., Varentsov, I.M., Golubev, N.G., andSokolova, E.Yu., Construction of Induction MagnetotelluricCurves from Profile Geomagnetic Data for the Study of Elec-trical Conductivity of the Continental Asthenosphere in theEMSLAB Experiment, Fiz. Zemli, 1997, no. 10, pp. 33–46.Vanyan, L.L., Varentsov, I.M., Golubev, N.G., andSokolova, E.Yu., Derivation of Simultaneous GeomagneticField Components from Tipper Arrays, Izvestiya, Physics ofthe Solid Earth, 1998, vol. 34, pp. 779–786.



Vanyan, L.L., Berdichevsky, M.N., Pushkarev, P.Yu., andRomanyuk, T.V., A Geoelectric Model of the Cascadia Sub-duction Zone, Izvestiya, Physics of the Solid Earth, 2002,vol. 38, pp. 816–845.Varentsov, I.M., A General Approach to the MagnetotelluricData Inversion in a Piecewise-Continuous Medium,Izvestiya, Physics of the Solid Earth, 2002, vol. 38,pp. 913−934.Varentsov, I.M., Golubev, N.G., Gordienko, V.V., andSokolova, E.Yu., Study of the Deep Geoelectric Structurealong the Lincoln Line (EMSLAB Experiment), Fiz. Zemli,1996, no. 4, pp. 124–144.Vozoff, K., The Magnetotelluric Method, ElectromagneticMethods in Applied Geophysics, Tulsa: SEG, 1991, vol. 2,pp. 641–711.Wannamaker, P.E., Stodt, J.A., and Rijo, L., A Stable FiniteElement Solution for Two-Dimensional Magnetotelluric

Modeling, Geophys. J. R. Astron. Soc., 1987, vol. 88,pp. 277–296.Wannamaker, P.E., Booker, J.R., Filloux, J.H., et al., Magne-totelluric Observations across the Juan de Fuca SubductionSystem in the EMSLAB Project, J. Geophys. Res., 1989a,vol. 94, no. B10, pp. 14 111–14 125.Wannamaker, P.E., Booker, J.R., Jones, A.G., et al., Resistiv-ity Cross Section through the Juan de Fuca Subduction Sys-tem and Its Tectonic Implications, J. Geophys. Res., 1989b,vol. 94, no. B10, pp. 14 127–14 144.Weaver, C.S. and Michaelson, C.A., Seismicity and Volcan-ism in the Pacific Northwest: Evidence for the Segmentationof the Juan de Fuca Plate, Geophys. Res. Lett., 1985, vol. 12,pp. 215–218.Zinger, B.Sh., Corrections for Distortions of MagnetotelluricFields: Limits of Validity and Static Approach, Surv. Geo-phys., 1992, vol. 57, pp. 603–622.

magnetovariational sounding: new possibilities

Documents