Magnetotelluric (MT) and seismic data, collected during the MELT experiment at the Southern East Pacific Rise (SEPR) 1,2 , constrain the distribution of melt beneath this mid-ocean-ridge spreading center and also the evolution of the oceanic lithosphere during its early cooling history. In this paper, we focus on structure imaged at distances ~100 to 350 km east of the ridge crest, corresponding to seafloor ages of ~1.3 to 4.5 Ma, where the seismic and electrical conductivity structure is nearly constant, independent of age. Beginning at a depth of about 60 km, there is a large increase in electrical conductivity and a change from isotropic to transversely anisotropic electrical structure with higher conductivity in the direction of fast propagation for seismic waves. Because conductive cooling models predict structure that increases in depth with age, extending to about 30 km at 4.5 Ma, we infer that the structure of young oceanic plates is instead

[1] The electromagnetic data from the Mantle Electromagnetic and Tomography (MELT) experiment are inverted for a two-dimensional transversely anisotropic conductivity structure that incorporates a correction for three-dimensional topographic effects on the magnetotelluric responses. The model space allows for different conductivity values in the along-strike, cross-strike, and vertical directions, along with imposed constraints of model smoothness and closeness among the three directions. Anisotropic models provide a slightly better fit to the data for a given level of model smoothness and are more consistent with other geophysical and laboratory data. The preferred anisotropic model displays a resistive uppermost 60-km-thick mantle independent of plate age, except in the vicinity of the ridge crest. In most inversions, a vertically aligned sheet-like conductor at the ridge crest is especially prominent in the vertical conductivity. Its presence suggests that the melt is more highly concentrated and connected in the vertical direction immediately beneath the rise axis. The melt zone is at least 100 km wide and is asymmetric, having a greater extent to the west. Off-axis, and to the east of the ridge, the mantle is more conductive in the direction of plate spreading at depths greater than 60 km. The flat resistive-conductive boundary at 60 km agrees well with the inferred depth of the dry solidus of peridotite, and the deeper conductive region is consistent with the preferred orientation of olivine inferred from seismic observations. This suggests that the uppermost 60 km represents the region of mantle that has undergone melting at the ridge and has been depleted of water (dissolved hydrogen). By contrast, the underlying mantle has retained a significant amount of water.

S U M M A R YRobust magnetotelluric response function estimators are now in standard use in electromagnetic induction research. Properly devised and applied, these have the ability to reduce the influence of unusual data (outliers) in the response (electric field) variables, but are often not sensitive to exceptional predictor (magnetic field) data, which are termed leverage points. A bounded influence estimator is described which simultaneously limits the influence of both outliers and leverage points, and has proven to consistently yield more reliable magnetotelluric response function estimates than conventional robust approaches. The bounded influence estimator combines a standard robust M-estimator with leverage weighting based on the statistics of the hat matrix diagonal, which is a standard statistical measure of unusual predictors. Further extensions to magnetotelluric data analysis are proposed, including a generalization of the remote reference method which utilizes multiple sites instead of a single one and a two-stage bounded influence estimator which effectively removes correlated noise in the local electric and magnetic field variables using one or more uncontaminated remote references. These developments are illustrated using a variety of magnetotelluric data.

The physics governing galvanic distortion of natural source electromagnetic induction measurements is reexamined beginning from first principles. The conditions under which a decomposition of measured magnetotelluric response tensors and magnetic transfer functions is applicable are described, and the form of the decomposition describing distortion of the electric and magnetic fields is derived directly from the integral equation defining the scattering of electric and magnetic fields by surface heterogeneities. The inclusion of magnetic field galvanic distortion leads to indeterminacy of the regional magnetotelluric response in the form of scaling by frequency‐dependent, complex factors controlled by two unknown real constants. This is a generalization of the well‐known static shift effect from electric field galvanic distortion and can in principal be removed if the magnitude and phase of the regional response are known at some frequency. Distortion of the magnetic transfer function is shown to be even more indeterminate, containing a term proportional to one of the regional magnetotelluric responses which is inseparably additive to the regional magnetic transfer function, as well as the complex scaling seen for magnetotellurics. A set of simultaneous nonlinear equations describing the full electric and magnetic field galvanic distortion decomposition of the magnetotelluric response tensor and magnetic transfer function is derived, and methods for their solution are described, including implementation of jackknife error estimates. The full magnetotelluric decomposition is applied to severely distorted data from the Canadian shield and seafloor data from the EMSLAB experiment. In both cases, magnetic field galvanic distortion is important at periods under a few thousand seconds. This suggests that greater attention to galvanic distortion of the magnetic field is needed during magnetotelluric surveys.

Robust estimation of power spectra, coherences, and transfer functions is investigated in the context of geophysical data processing. The methods described are frequency-domain extensions of current techniques from the statistical literature and are applicable in cases where section-averaging methods would be used with data that are contaminated by local nonstationarity or isolated outliers. The paper begins with a review of robust estimation theory, emphasizing statistical principles and the maximum likelihood or M-estimators. These are combined with section-averaging spectral techniques to obtain robust estimates of power spectra, coherences, and transfer functions in an automatic, data-adaptive fashion. Because robust methods implicitly identify abnormal data, methods for monitoring the statistical behavior of the estimation process using quantile-quantile plots are also discussed. The results are illustrated using a variety of examples from electromagnetic geophysics. INTRODUCTIONReliable estimation of power spectra for single data sequences or of transfer functions and coherences between multiple time series is of central importance in many areas of geophysics and engineering. While the effects of the underlying Gaussian distributional assumptions on such estimates are generally understood, the ability of a small fraction of non-Gaussian noise or localized nonstationarity to affect them is not. These phenomena can destroy conventional estimates, often in a manner that is difficult to detect.Problems with conventional (i.e., nonrobust) time series procedures arise because they are essentially copies of classical statistical procedures parameterized by frequency. Once Fourier transforms are taken, estimating a spectrum is the same process as computing a variance, and estimating a transfer function is a similar procedure to linear regression. Because these methods are based on the least squares or Gaussian maximum likelihood approaches to statistical inference, their advantages include simplicity and the optimality properties established by the Gauss- Paper number 5B5911 0148-0227/87/005B-5911505.00 residuals are drawn from a multivariate normal probability distribution, then the least squares result is also a maximum likelihood, fully efficient, minimum variance estimate. In practice, the regression model is rarely an accurate description due to departures of the data from the model requirements. Most data contain a small fraction of unusual observations or "outliers" that do not fit the model distribution or share the characteristics of the bulk of the sample. These can often be described by a probability distribution which has a nearly Gaussian shape in the center and tails which are heavier than would be expected for a normal one, or by mixtures of Gaussian distributions with different variances.Two forms of data outliers are common: point defects and local nonstationarity. Point defects are isolated outliers that exist independent of the structure of the process under study. In this paper the principles o...

The magnetotelluric method is a technique for imaging the electrical conductivity and structure of the Earth, from the near surface down to the 410 km transition zone and beyond. This book forms the first comprehensive overview of magnetotellurics from the salient physics and its mathematical representation, to practical implementation in the field, data processing, modeling and geological interpretation. Electromagnetic induction in 1-D, 2-D and 3-D media is explored, building from first principles, and with thorough coverage of the practical techniques of time series processing, distortion, numerical modeling and inversion. The fundamental principles are illustrated with a series of case histories describing geological applications. Technical issues, instrumentation and field practices are described for both land and marine surveys. This book provides a rigorous introduction to magnetotellurics for academic researchers and advanced students and will be of interest to industrial practitioners and geoscientists wanting to incorporate rock conductivity into their interpretations.

A resistivity cross section to depths exceeding 200 km has been derived from magnetotelluric observations along a profile near latitude 45øN from the Juan de Fuca spreading center, across the coastal subduction complex, the High Cascades volcanic arc, and into the back-arc Deschutes Basin region. In this two-dimensional interpretation, emphasis was placed on data approximating the transverse magnetic mode since these data are relatively robust to common departures from the two-dimensional assumption. The vertical magnetic field, however, has been very valuable in defining structure of the offshore sediments, of the oceanic asthenosphere and below the arc volcanics of the Westem and High Cascades. The transverse electric data on land suffer a variety of three-dimensional effects, making their interpretation very difficult. In contrast, the greater uniformity of upper crustal conditions on the seafloor allowed a good fit to both modes of the impedance plus the vertical magnetic field at least down to 104 s' Important components of . While a number of aspects of the subsurface resistivity can be deduced straight from inspection of the data, computer simulation of the observations helps to quantify more rigorously the permissible or required structures. In this paper, we derive a two-dimensional resistivity model, with a presumed northsouth strike, by trial-and-error fitting of our MT measurements with a finite element forward modeling algorithm. Constraints on model geometry from independent geological or geophysical investigations have been incorporated where justified. Our paper is divided into three major sections. First, we explain the approach toward two-dimensional modeling. On land, the transverse magnetic (TM) impedance functions are emphasized because theory and experience show that they are more robust to plausible three-dimensional effects in the region than are the vertical magnetic field or, especially, the transverse electric (TE) impedance. Upper crustal structure on the seafloor appears to be rhuch simpler than on land, however, and both modes of impedance plus the vertical field are fit fairly well. Second, the model cross section is described and its goodness of fit to the data demonstrated. Model uniqueness is investigated by perturbing certain features of the model and assessing the increase in misfit.

One‐dimensional electrical conductivity structure in the mid‐mantle of the one‐fourth of the Earth beneath the north Pacific Ocean was obtained by a semi‐global electromagnetic induction study. Electromagnetic response functions estimated from electric field variations measured by submarine cables and geomagnetic field variations obtained by magnetic observatories and long‐term observations sites were inverted into radially symmetric conductivity distribution by taking the distribution of land and ocean at the surface into account. As a most preferred model, a smooth conductivity‐depth profile was obtained with two abrupt increases that possibly correspond to the seismic discontinuities at 410 and 660 km.