Title: 6.2 The effect of the Groom-Baily decomposition to the deep MT soundings in the Great Hungarian Plain (corrected asthenospheric map) Author: A. Adam Geodetic and Geophysical Research Institute, H-9401 Sopron, POB 5, adam@ggki.hu More than 25 deep - broadband - magnetotelluric soundings (MTS) have been carried out in the Eastern part of Hungary in the Great Hungarian Plain. These MT soundings aimed first of all at the determination of the structure of the conducting asthenosphere besides the study of the surface sedimentary cover. As the greater part of the MTS curves are distorted by the basement relief their magnetotelluric structural anisotropy could generaly be high. Groom-Baily (GB) decomposition has been applied to reduce the 3D near-surface effect to decrease so the structural anisotropy. It resulted in the decrease of the MT anisotropy from 1.5 to 1.1 in a statistical sense. The GB distortion parameters - shear and twist - are in average between 7( and 11( (absolute values). The deviation of the asthenospheric maps calculated on the basis of the original 1D asthenospheric depth and those obtained after decomposition is relatively small. Besides the decomposition the galvanic distortions (static shift) should have been taken into account because the GB decomposition did not do it. As conclusion the paper gives a corrected asthenospheric map for the Great Hungarian Plain with average depth of 60 km in agreement with the early value of the author from 1963. Title: 6.3 INVESTIGATION OF ELECTRIC AND MAGNETIC GALVANIC DISTORTION USING A THREE-DIMENSIONAL NUMERICAL MODEL Authors: A.K. Agarwal and J.T. Weaver (Department of Physics & Astronomy and School of Earth & Ocean Sciences, University of Victoria, B.C., Canada, V8W 3P6. E-mails: numod@uvvm.uvic.ca and weaver@uvphys.phys.uvic.ca) Decomposition of the MT tensor when the electric field alone is subject to local distortion has been studied by various authors. The method has recently been extended to include magnetic field distortion. With the aid of a three-dimensional (3D) modelling program we investigate in this paper the validity of their assumptions about the nature of the electric and magnetic distortion matrices by computing the response of a synthetic model over a wide range of periods. The synthetic model consists of a 2D fault line plus a small 3D surface conductive anomaly in the shape of a cubical block near the fault and the periods for which its response is computed are T = 0.1, 1, 10, 50, 100, 500, 1000 and 5000 s. Our numerical study indicates that the inclusion of magnetic distortion is important and necessary for short periods below 10 s, but that the magnetic distortion parameters become complex at these periods so that the system of equations defining them is underdetermined and cannot be solved. On the other hand, the influence of magnetic distortion becomes insignificant for periods above 10 s where the usual assumption that the distorting matrix is real and frequency-independent is indeed approximately valid. At these longer periods the phases of the regional impedances and the ratios of electric distortion parameters are correctly recovered whether magnetic distortion is present or not. However, the diagonal elements of the magnetic distortion matrix are generally poorly resolved because the imaginary parts of the electric and magnetic distortion elements are completely ignored. It is concluded that, at least for this model, inclusion of magnetic distortion in decompositions of the MT tensor does not in general offer any improvement over the conventional decompositions in which only electric field distortion is taken into account. Title: 6.4 DOUBLE IP-EFFECT IN ELECTROMAGNETIC TRANSIENTS Authors: Barsukov P.O., Fainberg E.B. Institute of geoelectromagnetic investigations of Russian Academy of Sciences, 142092, Troitsk, Moscow region, fain@geo.igemi.troitsk.ru Well-known IP-effect, stipulated by propagation of electromagnetic field in frequency-dispersive media is displayed as inversion of sign of transient processes when TEM (Transient electromagnetic) measurements. Application of conventional inversion algorithms in this case is impossible. Development of algorithm accounting the nature of IP-effect is actual problem. Presented paper is devoted to investigation of the nature of IP-effect in granite massifs and their wind-blown crust. Numerous TEM soundings using single loop modification (the same transmitter and receiver antenna) showed that displayed IP-effect can contain at least two parts: subsurface and deep. Subsurface IP-effect displayed at early times of the transient processes is decreased sharply with increasing of antenna's size or when uplifting above of the earth surface. IP-effect, which is displayed at more late stages of transient processes, does not depend on uplifting of antenna; its value is increased proportionally to cubic degree of antenna's size. Our investigations showed that the signals in the range of negative sign are unstable in time. To investigate this phenomenon we carried out monitoring of IP-effect using fixed 50 m x 50 m antenna and TEM-FAST instrument operating in automatic regime. Subsurface IP-effect is most intensive in nighttime; the deep effect is distributed in time regularly. Behavior of subsurface IP-effect has correlation with variations of humidity in area of measurements. Probably, its nature is connected with the processes running in the system antenna-subsurface humid soil. Deep IP-effect and its time variations are stipulated by dynamics of hydrogeological state of buried rocks. The research were supported by Russian Foundation for Basic Research, project 97-05-64338 Title: 6.5 ON SENSITIVITY AND ROBUSTNESS 0F THE TM AND TE MODES OF THE TWO-DIMENSIONAL MAGNETOTELLURIC FIELD: PRINCIPLE OF INFORMATION COMPLEMENTARITY Authors: M.N.Berdichevsky, V.I.Dmitriev, E.E.Pozdnjakova Moscow State University, Moscow, Russia, E-mail: berd@geo.geol.msu.su The 2D modelling shows that the TM mode is more sensitive to near-surface resistive structures and to the lithosphere resistance, but the TE mode is more sensitive to deep conductive structures over- lapped by unbroken resistive layers. Again, the TM mode may have a rather high sensitivity to conductive faults which connect near-surface and deep conductors and form closed current circuits. Another important point is the robustness of the TM and TE modes to 3D effects. Wannamaker, Hohman and Ward showed that the in the model with a conductive prism the TM impedance is more robust to 3D effects than the TE impedance is. But in the model with a resistive prism the TM and TE impedances change places: here the TE impedance may be more robust than the TM impedance is. So, the traditional idea of higher robustness of the TM mode should be corrected. The proper wording is: " The TM mode is more robust to 3D effects caused by conductive structures, but the TE mode is more robust to 3D effects caused by resistive structures". Thus, the TM and TE modes satisfy the principle of information complementarity: the gaps left by one mode can be filled by another mode. Title: 6.6 DEPRESSION OF STATIC SHIFT IN MT-OBSERVATIONS BY A NORMALIZING TECHNIQUE WITH THE USE OF THE S-MEAN VALUES Authors: Feldman, I.S., Okulessky, B.A. ( Associated Institute for High Temperature, RAS, Izhorskaya str., 13/19, Moscow, 124412, Russia) Yakovlev, A.G. and Kulikov, V.V. (Lomonosov Moscow State University, Moscow, 119899, Russia) The influence of the static shift in MT-observations, or S-effect in the terminology of M.N. Berdichevsky, manifests itself just about everywhere. One of techniques of depression of this effect is based on the reduction (normalizing) of MTS-curves to the mean conductivity of deposits overlapping screening layers (Berdichevsky, 1988). The capabilities and limitations of this technique are considered with the help of 2-D and 3-D models with variable conductivity of formations overlapping screens. Both analytical estimations and results of finite-difference modelling were used. The influence of one and two screens are analyzed at various characteristic dimensions of conductivity irregularities and different galvanic constants at various levels. Several typical examples of the use of the normalizing technique at the real field data obtained in various region of Russia are presented. In all cases, high quality field data at high density of observations were used. These data were accompanied by known geological information about every area studied (TEM, reflection seismics, bore-hole logging). A scheme of the determination of statistic characteristics of the static shift from real data as well as zoning of these characteristics in the region are considered. Title: 6.7 A superposition of near surface and middle crustal galvanic distortion in long period MT-data. Authors: Mark Hoenig, Joerg Leibecker, Karsten Bahr Institute of Geophysics Herzberger Landstrasse 180 37083 Goettingen email: mark@willi.uni-geophys.gwdg.de Germany In August 1997 a field experiment in the german eifel was carried out, where 23 long period MT-stations were set up at an array of 90 x 110 km. The near surface lateral conductivity contrasts cause a galvanic scattering of the transfer functions over the whole period range (10 s - 4000 s). For very long periods with penetration depths ranging into the upper mantle the anisotropic middle crust also acts as a scatterer, as seen from data and modelling results. The aim of this work is to use the mid crust as a scale to extract the static shift caused by the near surface structures from the data. Then to model the mid crust, and thus be able to recognize the galvanic effects due to this layer, to finally get undistorted sounding curves at long periods by subtracting both distortions from the data. Title: 6.8 "Charting MT tensor decomposition in Mohr space" Author: F.E.M.(Ted) Lilley Research School of Earth Sciences Australian National University Canberra, ACT 0200, Australia email: Ted.Lilley@anu.edu.au Session: Distortion and dispersion effects: The Mark Berdicevski Symposium. oral/poster: Oral Tensor decomposition is presented as determining values for the most informative set of invariants of the tensor. These values are evident (along with other parameters involved in tensor decomposition) on Mohr diagrams for the tensor, when real and quadrature parts of the tensor are taken separately. There is benefit in being able to actually see, on a diagram, the parameters involved in a tensor decomposition process; and to see the decomposition taking place, step by step. The following set of invariants demonstrates the idea: Two values for the magnitudes of the real and quadrature parts of the tensor; Two values for the 2-D splits or anisotropies of the real and quad parts; Two values for the twists or skews, from their closest 2-D cases, of the real and quad parts; One value for the misalignment of the real and quad closest 2-D cases; One value for the geographic strike of the real closest 2-D case. The analysis also leads to the conclusion that tensor data should obey the inequality Zxy.Zyx < Zxx.Zyy for both real and quadrature parts taken separately. This criterion is most likely to fail (and so be useful) for the lower impedance mode of a highly anisotropic tensor. Data which fail this criterion are probably, in their lower impedance mode, indistinguishable from zero. Title: 6.9 Separation of static effects from the MT impedance by using the geomagnetic transfer functions. (II). Model calculation Authors: Hiroshi MUNEKANE and Hisashi UTADA Earthquake Research Institute, University of Tokyo Yayoi 1-1-1, Bunkyo-ku, Tokyo 113-0032, JAPAN munekane@eri.u-tokyo.ac.jp, utada@utada-sun.eri.u-tokyo.ac.jp Session: 6. Distortion and dispersion effects : The Mark Berdicevski Symposium Presentation: Oral (preferred) or poster Utada and Munekane (in this session) proposed a new method to remove static distortion effects from the MT impedance using an identity between the geomagnetic transfer functions and the spatial derivatives of impedance and horizontal magnetic components. This method has a theoretical advantage that makes decomposition possible even if the regional undistorted structure is three-dimensional. In this paper, a method is actually implemented and its validity is tested, especially in the case that regional undistorted structure is three-dimensional, by using synthetic data sets. First, regional two- and three-dimensional MT impedance are calculated at each given site. Then real distortion tensor is randomly chosen and synthetic distorted response is calculated at each site. Finally distortion tensor at each site is determined so as to minimize the gap from the given identity. A nonlinear least squares method (the Levenberg-Marquardt method) was applied for the determination. For simplicity, the gradients of horizontal magnetic components are assumed to be small as is in usual MT cases. Validity of this assumption and form of the object function are also examined through the model calculation. Title: 6.10 " Distortion effect of elongated surface conductivity-inhomogeneities on CSAMT measurements" Authors: G. Petho(1) and L. Ficsor(2) (1) University of Miskolc Dept. of Geophysics (2) University of Miskolc Dept. of Information Technology Address: H-3515 Miskolc, Egyetemvaros, Hungary e-mail: (1): gfpg@gold.uni-miskolc.hu (2): ficsor@zeus.iit.uni-miskolc.hu Session: Distortion and dispersion effects- the Mark Berdichevsky Symposium The paper is attended to be oral presentation. The paper presents the comparison of the frequency domain EM responses of two-dimensional surface-conductivity inhomogeneities to a horizontal electric dipole source to be parallel and perpendicular to the structural strike. The EM responses are determined along an in-line profile perpendicular to the strike at the frequency range of 100Hz-0.1Hz. The applied numerical procedure is a finite-difference method assuming 2.5-D situation. After solving the partial differential equation system in the along-strike wavenumber domain the spatial EM field components are determined by a numerical inverse Fourier transform. It was experienced that 27 along-strike wavenumbers and three different grid geometries were needed to get solution. The shift of electric field component amplitude of dipole axial and dipole equatorial array can be observed in a different degree in the function of frequency over the inhomogeneities with constant phase values at low frequencies. Some similarities between EM responses of 2D MT HPOL and dipole axial array and those of 2D MT EPOL and dipole equatorial array can be experienced. However, in practical cases CSAMT measurements are not carried out in the plane wave zone, rather they are made in the part of transition zone close to the plane wave zone resulting in the coupling of the two source polarizations. Session: 8, oral Title: 6.11 Distortion Correction in a Very Conductive Medium: Comparison of tensor decomposition and modeling Authors: P Tarits, A.D. Chave, and R Nolasco Seafloor MT data from rough bathymetric environments can be significantly distorted, and both the electric and especially the magnetic fields may be affected. We present examples of data from both the MELT experiment and an offshore survey near Tahiti for which the classical galvanic distortion removal technique (tensor decomposition) either works or fails. We will compare the undistorted transfer functions to both the decomposed transfer functions and those corrected using a thin sheet algorithm that explicitly takes into account the inhomogeneous bathymetry-induced variability of the sea conductance. We can then test the relative weight of the galvanic and inductive parts of the distorted fields. Pascal Tarits, Rita Nolasco tarits@univ-brest.fr maria@univ-brest.fr UBO/IUEM, UMR CNRS 'Domaines Oceaniques' Place Nicolas Copernic, F-29280 Plouzane, France Alan Chave achave@whoi.edu Woods Hole Oceanography Institute Woods Hole, MA02243, USA session 6 Title: 6.12 Separation of static effects from the MT impedance by using the geoma gnetic transfer functions. I. Theoretical background Authors: Hisashi UTADA and Hiroshi MUNEKANE Earthquake Research Institute, University of Tokyo Yayoi 1-1-1, Bunkyo-ku, Tokyo 113-0032, JAPAN e-mail: utada@utada-sun.eri.u-tokyo.ac.jp, munekane@eri.u-tokyo.ac.jp Session: 6. Distortion and dispersion effects : The Mark Berdicevski Symposium Presentation: Oral (preferred) or poster The Groom-Bailey decomposition is widely used to remove distortion effects from the MT impedance. Distortion may be expressed by a simple real tensor operated to the impedance in general. The problem is how to solve the linear equation to obtain the distortion tensor and then undistorted impedance. The Groom-Bailey's method assumes a two- dimensionality for the undistorted impedance. In this paper we propose a method which uses a relationship between the spatial derivatives of the horizontal electric field and the vertical geomagnetic component, which can be directly derived from Faraday's low of induction. This relation leads us to an identity between the geomagnetic transfer functions and the spatial derivatives of the impedance and the horizontal magnetic components. If the horizontal magnetic field is nearly uniform as is in usual cases, this relation directly indicates that the transfer function means the gradient of the impedance (or, in other words, induction vectors point to the conductors). The present method determines the distortion real tensor so that this relation is satisfied and therefore has two advantages: (1) it clearly separates static effects from the impedance tensor, and (2) it makes decomposition possible even if the regional undistorted impedance is three-dimensional. Title: 6.13 CHARACTERIZATION OF THE MAGNETOTELLURIC TENSOR IN TERMS OF ITS ROTATIONAL INVARIANTS WITH PHYSICAL INTERPRETATIONS. Authors: J. T. Weaver, A. K. Agarwal (Department of Physics & Astronomy and School of Earth & Ocean Sciences, University of Victoria, Victoria, B.C., Canada, V8W 2Y2) and F. E. M. Lilley (Research School of Earth Sciences, Australian National University, Canberra, 0200 ACT, Australia). At the 13th Workshop in Onuma, Szarka & Menvielle presented an analysis of the rotational invariants of the impedance tensor and showed that the maximum number of mutually independent invariants was seven. We have characterized the MT tensor by seven such invariants which differ somewhat from those proposed by Szarka & Menvielle, and we have given a physical interpretation of the vanishing of each one. Two provide information for a 1D model, the second pair are associated with 2D anisotropy, the third pair provide information on possible galvanic distortion, while the seventh invariant is associated with the presence of a 3D structure and is closely related to Bahr's ``phase sensitive skew''. Each one is also identified on a Mohr circle diagram where its invariant property becomes geometrically obvious. In order to test the physical interpretation of each invariant we have constructed a 3D numerical model comprising a 2D fault line and a small conductive cubical anomaly at the surface of the model near the fault. The regional field is taken to be inclined at an angle of 40 degrees to the fault. The complex elements of the MT tensor have been computed for several frequencies at various points in the model and the seven invariants have been evaluated. The physical interpretations of the invariants have been verified for this simple model and the direction of the strike of the fault line has been recovered with acceptable accuracy in those cases where it is possible to do so. Title: 6.14p DECOMPOSITION WITH A GUI INTERFACE Authors: Julie Allen and Alan Chave (Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA; jallen@whoi.edu or alan@whoi.edu) A graphical user interface (GUI) called MagnetoTelluric Interactive Processing System (MTIPS) has been implemented as a "front end" for the widely circulated RRRMT robust controlled leverage MT processing and GBDECOMP electric/magnetic galvanic distortion decomposition programs. The interface allows the user to enter and modify the input parameters to either program using a graphical form. It also implements complete graphical display, including hardcopy, of all output parameters from the programs, including the original and decomposed MT responses, statistical information about these quantities, and color-coded time series plots to show which portions of the data have been retained or rejected during processing. The MTIPS GUI is written in public domain Tcl/Tk (Tcl version 7.5/Tk version 4.1) using the BLT 2.1 graph extension and Embedded Tk 1.7. MTIPS can be run on any Unix platform with Tcl/Tk (and the BLT extension) installed, or it can be compiled into a stand-alone executable that will run on any binary-compatible computer, even if Tcl/Tk is not installed. Examples illustrating the convenience of MTIPS data processing will be shown. Title: 6.15p THE NATURE OF CRUSTAL ANOMALY OF ELECTRICAL CONDUCTIVITY OF BYELORUSSIAN ANTICLINE (session 9) Author: Astapenko V. N. (Institute of geological sciences of National Academy of Sciences of Byelorussia, 220600, Minsk, Zhodinskaya 18, geology@ns.igs.ac.by) Fainberg E. B. Institute of geoelectromagnetic investigations of Russian Academy of Sciences, 142092, Troitsk, Moscow region, fain@geo.igemi.troitsk.ru) Deep magnetotelluric soundings along 550 km profile Slonim-Grodno-Vitebsk crossing Byelorussian-Baltic granulite belt (East-European platform) have been fulfilled in the frame of International Europrobe project. Technology of synchronous measurements in base and field stations was applied. MT and MV curves have been constructed in 96 sites. MT transfer functions of deep sounding were constructed in the frequency range 0.0003-3000 Hz. They demonstrated clearly the existence of crustal conductive layer practically everywhere along the profile. In addition, in several sites local conductive anomalies in upper part of basement were found. For understanding of the nature of this layer it was important to find precisely the depth of the layer. For that, exhaustive analysis of the data based on thin sheet modeling, application of dynamic correction of transfer functions to account distorting influence of subsurface inhomogeneities was carried out. The final results of modeling, correction and inversion permitted to locate the conductive layer in the crust at depth 10-12 km depth. Specific resistance and conductance of the layer is of order of 300 Ohmm and 30-50 S correspondingly. Resistance of containing rocks is two orders higher. It should be noted that the location of the layer coincides with the layer of velocity 6.3 km/s. For explanation of a nature of this layer, data of laboratory measurements of rocks, characteristic for the region under investigations produced in Ukraine have been used. Analysis showed that essential part of rocks of crystalline basement of granulite belt consists of amphibole minerals. Modeling of behavior of amphibolites at high temperature and pressure when in lab. conditions was fulfilled in Lebedev et. al. (1993). In their experiment PT-conditions up to 30 km simulated gradual deepening of rocks with depth. Acoustic, Roentgen-structural and thermal measurements have revealed existence of high anisotropy and formation of numerous cracks with a maximum at T~245 oC and P~270 MPa. Such PT values are characteristic of depth at 12 km. The water and tangential tensions essentially strengthened this processes. The important question is the source of the water, existence of which is necessary for explanation of this process. The entering of a water on these depths can goes at the side and then along horizontal cracks (meteoric water) and from below along deep faults (mantle).. According to Glover and Vine (1995), 0.1% increase of the water contents on this depth is enough for explanation of observable jump of conductivity. This research was supported by Russian Foundation for Basic Research, project 96-05-65830 Title: 6.16p ELECTROMAGNETIC RESONANCE IN TRANSIENTS? Authors: Barsukov P.O., Fainberg E.B. Institute of geoelectromagnetic investigations of Russian Academy of Sciences, 142092, Troitsk, Moscow region, fain@geo.igemi.troitsk.ru (session 6) When propagation of EM pulses in real geological medium, the energy of a magnetic field is accumulated in induction currents, while electrical charges - on borders of inhomogeneities. Change of energy, necessary for existence of oscillatory character of the processes is possible only if the energy of the processes is commensurable. However, in real geological conditions energy of electrical charges much less of magnetic ones, mainly, because of low values of permittivity and high values of conductivity of rocks. Theoretical and the experimental research of mountain rocks confirm this conclusion. However in a nature there are possible situations in which resonance fluctuations do observed. One example is TEM research carried out in mountain regions of Tyan-Shang. For soundings the scheme of measurements using one-loop system and DC soundings of various configuration have been used. Research of resonance effects revealed that period of fluctuations does not depend on the size and form of transmitter and receiver and height of the antenna above ground. The period of fluctuations don't depend also on type of electrical excitation and reception of a field in an interval of times 20-50 ms. Frequency of fluctuations decreased when increasing of width and depth of the gorge. The amplitude of oscillatory processes increased in central part of the gorge, where decrement of attenuation also decreased. Laboratory research of samples of these rocks, taken from barrens, has shown abnormal high values of permittivity and polarizability in a range up to 100 kHz. Resonance processes in the samples of rocks did not register. The possible nature of this phenomenon is discussed. The research were supported by Russian Foundation for Basic Research, project 97-05-64338 Title: 6.17p IN 3D MODELS WE CAN FACE THE VIOLATION OF DISPERSION RELATION Authors: M.N.Berdichevsky, D.O.Pokhotelov Moscow State Unoversity, Moscow, Russia E-mail: berd@geo.geol.msu.su We examined a model with a local 3D inclusion imposed on a regional 2D structure (the Bahr model).Here the 3D impedance is a linear combination of both 2D impedances and at certain direction of x, y axes has zero in the upper frequency half plane (but no poles). Calculation shows that the dispersion relation between real and ima- ginary parts of the 3D impedance is preserved here but the relation between apparent resistivities and phases is drastically violated. This result resolves a long discussion about universality of the dispersion relations. Abstract for session > Distortion and dispersion effects: The Mark Berdichevsky Symposium Title: 6.18p Full Galvanic Distortion Decomposition of CSAMT data. Application to Buchans Mine Line 14. Authors: Xavier Garcia, University of Barcelona David Boerner, Geologicla Survey of Canada Using a 1D inversion algorithm based on the full galvanic decomposition of the electromagnetic fields, we have inverted 10 stations recorded in Buchans Mine, Newfoundland. Buchans Mine is situated over one of the biggest sulfite deposits of the world, presenting a highly resistive crystalline basement. Previous works in the area suggested that the difference in the surface resistivity amongst stations were due to the existence of fractures and hematitic zones, producing a high distortion of the CSAMT responses. The inversion code used is based in the Marquardt-Levenbergs method, using a earth model consisting in a layered model distorted through three distortion matrices, the regional model and the distortion parameters have been obtained. The results show a three layered model that correlates with the seismic profile interpretation, giving important information about two thrust packages. The study of the distortion parameters obtained from the inversion suggest that the galvanic magnetic distortion should be take into account in order to interpret these data, since the magnetic distortion parameters do not vanish at any frequency. TITLE : 6.19p Distortion Analysis of magnetotelluric data in the Bolivian Altiplano considering the inherent errors of the measured data. AUTHORS: Pamela Lezaeta, Freie Universitaet Berlin, FR Geophysik, Maltesertsr. 74-100, D-12249, Berlin, Germany. Email: pamela@geophysik.fu-berlin.de Patrick Denny, GFZ-Potsdam, Telegrafenberg, D-14473 Potsdam, Germany. Email: denny@gfz-potsdam.de ABSTRACT: A Gaussian distribution with a 95% confidence limit is used for the spectral analysis of MT data measured in the Bolivian Altiplano, after processing with Egbert (1998) code with a remote reference station. The statistical distribution and respective confidence limit of each tensor element is estimated. With an a priori distribution of the tensor elements the distribution of any impedance based parameter can be determined. Using first order Taylor approximation one calculates the error of the parameter function of the tensor elements, if these lasts are independent variables with no large errors. The propagation errors of several parameters for MT sites located in the Bolivian Altiplano are determined. The Bahr (1991) telluric distortion type is obtained with its probability distribution and strike angles are compared with those from Chave's (1993) decomposition code, giving the strike of the probable two-dimensional structure. A different statistical approach is also implemented which considers the errors in the measured quantities to be Gaussian after the removal of extreme outliers. A confidence limit is derived for the apparent resistivity, creating a new expression for error bars based on the quotient of Gaussian distributions inherent in the tensor elements using this treatment. Title: 6.20p LATITUDINAL VARIATION OF GEOMAGNETIC FIELDS AT THE BRAZILIAN EQUATORIAL ZONE AND ITS EFFECT ON MT SURVEYS* Author: Antonio L. Padilha (INPE, C.P. 515, 12201-970 Sao Jose dos Campos, Brazil; e-mail: padilha@dge.inpe.br) In a previous paper, Padilha et al. (Geophys. Res. Lett., V.24: pp.89, 1997) have shown that the equatorial electrojet (EEJ) currents did not affect significantly MT soundings carried out across the magnetic dip equator at the Brazilian equatorial zone. By using geomagnetic variation signals measured simultaneously to the MT experiment at a chain of ground-based equatorial and mid-latitude stations, an attempt is done to explain those results in terms of the behaviour of the primary inducting field during the survey. It is observed that the amplitude of the geomagnetic variations appears horizontally homogeneous within the study area (from -3 deg. to +3 deg., geomagnetic latitude), without any sharp change under the dip equator. It can be concluded that the primary field in the analysed frequency range may be considered sufficiently uniform in the horizontal direction thus satisfying the Tikhonov-Cagniard plane-wave criterion. The same geomagnetic data also shows that, if exists any EEJ source effect, it would be restricted to the transition zone (between 3 deg. and 5 deg., at both sides of the dip equator) and to frequencies lower than 0.1 mHz. * This work was supported by a FAPESP Grant No 93/0652-2. Title: 6.21p Dimensionality analysis and 2D inversion of MT data from the Hessian Rift, Germany Authors: Palshin N.A (1)., Tezkan B.(2) and Vanyan L.L(1) 1) Shirshov Institue of Oceanolgy, Russian Academy of Sciences, Russia 2) Institute of Geophysics and Meteorology, University of Cologne, Germany The measurements were made at more than twenty sites in Hessian Rift area (Germany) during the last decade (Flushe, 1984; Tezkan et al.,1988; Peter, 1993). The Hessian Rift conductivity structure is characterized by presence of linear high conductive zones and 3D nearsurface inhogeneties. The impedance and geomagnetic transfer functions were obtained in a wide period range from 10 sec to 16 hours. The aim of the study was to to analyze the distortion effects and to derive background conductivity structure as a result. For the purposes of our study the effect local 3D near surface inhomogeneties (static shift) could be treated as geological noise. To resolve the static shift we need to quantify the dimensionality of the dominant conductivity structure. The comparison of different techniques of dimensionality analysis (Swift's diagonal minimization, phase-sensitive Bahr's method and GB decomposition) was carried out. Reference GDS data was used to estimated regional static shift. Regional 1D (background) conductivity structure was derived from the data set. GB decomposed MT data (after resolving the static shift) together with geomagnetic transfer functions was used to construct a consistent 2D conductivity model using 2D OCCAM. Title: 6.22p 3D MODELLING STUDIES OF MAGNETIC DISTORTION EFFECTS IN GDS DATA Authors: Patricia Ritter GeoForschungsZentrum Potsdam, Telegrafenberg, D-14473 Potsdam, Germany pritter@gfz-potsdam.de The magnetic response of a confined structure depends on its scale length, its depth, its dimension and its conductivity contrast with respect to the host material. These parameters determine the period range in which anomalous magnetic fields are generated by currents induced inside a conductive anomaly. At longer periods, the response is produced increasingly by currents induced in the host material. These currents, however, may be concentrated close to and inside the anomalous body. The 3D modelling studies show that these two processes can be distinguished in the magnetic response function by examining its phase. It changes from negative values in the period range of local induction to positive values, approaching the regional impedance phase, in the range where current deviation is the predominant process. The magnetic fields created by the latter process may cause a misleading interpretation of induction arrows especially in terms of regional strike directions. A detailed study of anomalies of different sizes and conductivities embedded in 1D or 2D background structures shows how these magnetic distortion effects may be recognized in GDS data. Title: 6.23p The effect of near-field artificial electromagnetic fields on Magnetotelluric parameters. Author: G. Santarato (1) and P. Palangio (2) (1) Istituto di Mineralogia - Ferrara, Italy (2) Istituto Nazionale di Geofisica - Rome, Italy Five-components magnetotelluric (MT) measurements were recorded in a site of Central Italy. In the frequency range 0.1 Hz to 0.001 Hz, in the presence of an artificial field due to a small Italian d.c. powered railway, which runs some 20 km apart from the measuring site, MT impedance and tipper were determined. Since the railway was switched on only during certain periods, the analysis was undertaken both in the presence and absence of the artificial field. The main effects of this field resulted of near-field type, which can often be recognised on the estimated parameters: abnormally high values of estimated apparent resistivities, while accompanied by a small variance, high slopes on apparent resistivity curves, very small values of impedance phases. All these effects cannot be separated "a posteriori" from the natural signal impedance response. Since the railway magnetic field is prevailingly vertical, its effects are particularly strong on the tipper, which fairly exceeds 1, while no earth structure exists where tipper can reach 1. An analysis of the problem is shown as a function of the natural signal strength and a procedure is suggested to avoid the artificial field distortion. Title: 6.24p TRANSIENT FIELD SOLUTION FOR THIN LAYER OF A FINITE THICKNESS Authors: B. Sh. Singer and A. Green B. Sh. Singer (presenting author), CRCAMET, Bldg.E5A, Macquarie University, North Ryde, NSW 2109, Australia, Tel.:(61-2)9850-9281, Fax.:(61-2)9850-8366, E-mail: bsinger@laurel.ocs.mq.edu.au A. Green, CRCAMET, Bldg.E5A, Macquarie University, North Ryde, NSW 2109, Australia, The transient solution found by J. C. Maxwell for a thin conductive layer is still widely used for interpretation of airborne electromagnetic data. When interpreting the data acquired over a conductive (rigolith) layer that in practice has a finite thickness, the solution is valid only on the late stage of the decay of the currents induced in the layer. The only parameter that can be found in such an interpretation is the integrated conductance of the layer. We derive a solution that retains the simplicity of the classic solution by J. C. Maxwell. In addition to the layer conductance, the new solution depends on the layer thickness. From our numeric experiments, when applied to a conductive layer of a finite thickness, the new solution becomes applicable at a delay that is by several orders of magnitude shorter than that of the thin sheet solution. Thus, the solution allows an interpreter to estimate the thickness as well as the conductivity of the subsurface conductive layer. Title: 6.25p SEPARATION OF THE REGIONAL AND LOCAL ASPECTS FROM THE GDS DATA Authors: V. Steflea, A. Soare, G. Cucu The transient geomagnetic field penetrating the Earth suffers an attenuation and phase change, differentiated on deep levels, depending on the variation spectrum and conductivity distribution in the subsoil. For the geomagnetic depth sounding method (GDS) the transfer function frequently used are the induction vectors results from a linear relation between the variational magnetic components. The orientation and the inclination of the inductor vector can be indicator for the existence and the geometry of a conductivity contrast in the subsoil. The Romanian territory has been covered by a relatively dense network where the induction vectors present features which can be correlated with some major geological elements. In order to try the separation of effects produced by different regional and local causes we represent in the form of maps the separate distribution of a and b values of the components of induction Wiese vector. These maps are analysed by polynomial approximations of different orders which coefficients Ki are determined and is established the surface which fits better the distribution of data. The images of different orders offer information about the position and the depth of the anomalous sources. This kind of analysis was established for all Romanian territory and for selected zones (especially for Dobroudja) and the obtained images indicate well the most important faults and structural blocks.