GP14A-01 INVITED 16:00h
The use of 3D Magnetotelluric Inversion for Exploration in Complex Geologic Environments: Potential Pitfalls and Real World Examples
The interpretation of magnetotelluric data collected in complex geological environments requires not only that the data themselves are reliable, but that the interpretation schemes are reliable. Unfortunately, since magnetotellurics is a diffusion phenomenon rather than wave propagation, it is difficult (if not impossible) to present MT data in an intermediate form corresponding to a seismic time-section. This also has implications for imaging schemes due to the inherent inability to resolve sharp boundaries. This does not mean that you can not find a sharp boundary model consistent with the data, but rather that they are NOT required by the data. The interpretation process, including inversion, is clearly subjective (for example, what exactly does 'smooth' mean?). In this short talk, we will examine several issues inextricably linked to interpretation of magnetotelluric data in complex geological environments including: (1) how the assumptions inherent in the inversion process restrict or distort the results, (2) what are the physical limits of our interpretations, and, (3) what are the sensitivity of the data to the geology. These issues will be addressed using both synthetic and real field data from a variety of complex exploration targets including oil exploration in over-thrust and sub-volcanic regions.
GP14A-02 16:30h
A Non-Linear Inversion for the Global 3-D Electrical Conductivity Distribution in the Upper to Mid-Mantle
The case for substantial heterogeneity in mantle conductivity has stimulated the development of methods for solving Maxwell's equations in a heterogeneous conducting sphere. A global 3-D frequency domain forward solver has been devised (Uyeshima & Schultz, 2000), accurate and efficient enough to be an attractive kernel of a practical inverse method. The solver employs a staggered-grid finite difference formulation in spherical coordinates. The induced fields are found as a solution to the integral form of Maxwell's equations, while the system is solved using stabilised biconjugate gradient methods. A single, accurate forward solution takes approx. 4 minutes on 5 GFLOP (peak) processor. The aim of our present research is to produce an inverse solver, to be applied to the Fujii & Schultz (2002) data set of globally-distributed EM response functions, which would reconstruct the 3-D electrical conductivity distribution in the upper to mid-mantle. Geophysical inversion is an ill-posed problem, therefore the aim is to apply suitable parameter constraints and a nonlinear search algorithm to identify candidate minima, then to apply local gradient methods around those minima. Our specific target involves designing a fast enough global optimisation routine that would allow us to produce at least one fully 3-D starting model, optimal with respect to the RMS misfit between the data and the forward solutions. A new and very flexible inverse solver has been developed utilizing parallel optimisation routines to obtain a starting model that satisfies the data. 3-D simulations have been run, the parametrization based on a spherical harmonic representation of a chess board model of varying degree and order. The inversion has demonstrated accurate fidelity in reproducing resolvable features of the test model. A study has been made of the reduction in fidelity as the number and distribution of observatory sites on the Earth's surface is degraded. An inversion of the Fujii & Schultz (2002) geomagnetic data set is underway. We also discuss implementing a linearised sensitivity analysis as part of the inversion.
GP14A-03 16:45h
Three-dimensional MT Modeling and Inversion for Offshore Petroleum Exploration
There is growing interest in the application of marine magnetotelluric (MT) methods for offshore petroleum exploration (Constable et al., 1998; Hoversten et al., 2000). However, until recently, the interpretation of the sea-bottom MT data was performed mostly by two-dimensional (2-D) modeling and inversion methods (de Groot Hedlin and Constable, 1990, 2004). At the same time, the typical sea-bottom geological structures, such as salt domes or petroleum reservoirs, are three-dimensional (3-D). In this paper, we use for 3-D marine MT inversion a new algorithm, which is based on an application of the quasi-analytical (QA) approximation in the initial stage of the iterative inversion, and of the rigorous integral equation forward modeling in the final stage of the inversion. We present the results of numerical modeling and inversion for the typical models of offshore sea-bottom petroleum reservoir in the presence of a salt dome structure. We also consider a case history of 3-D inversion of sea-bottom MT data collected by the Scripps Institution of Oceanography in Gemini Prospect, Gulf of Mexico. The obtained 3-D inverse model correlates reasonably well with the location and shape of the salt dome structure known from the seismic exploration data. This result demonstrates that 3-D inversion method can be used for interpretation of the sea-bottom MT data for offshore exploration.
GP14A-04 17:00h
Development of Nested Modeling in the Modular Electromagnetic Modeling and Inversion Software (MEMIS) System: A New Software Tool for EARTHSCOPE
We are developing modular electromagnetic (EM) modeling and inversion software (MEMIS) as a tool for novel modeling and hypothesis testing, and for prototyping inversion algorithms for electromagnetic data. Flexibility in the design facilitates incorporating new approaches to modeling in geologically complex, multi-scale environments. As an illustrative example we consider development of a nested modeling capability, with boundary conditions for a smaller target area determined from a regional scale geoelectrical model solution. In addition to providing more realistic boundary conditions, our approach allows quantitative assessment of the sensitivity of data to uncertainties in large-scale boundary conditions. This nested approach to EM data interpretation will allow information about the continental scale geoelectric environment obtained from the EARTHSCOPE big-foot magnetotelluric (MT) array to be integrated into interpretations of higher resolution local surveys. As an initial test example, we consider MT imaging of the San Andreas Fault in the vicinity of SAFOD at Parkfield (California), an area where both high resolution and larger scale MT surveys have been conducted, or are being planned. Through forward modeling exercises we explore the interaction of far-field features, such as electrical currents induced in the Pacific Ocean, with the local three-dimensional fault structure. Data sensitivities are used as an initial assessment of the potential impact of improved regional scale models on interpretation of high-resolution MT surveys in the vicinity of SAFOD. A long-term goal is to evolve MEMIS into a flexible, extensible open source system for EM modeling and inversion that will allow the broader earth science community to make better use of the multi-scale EM datasets expected to be collected through national and international programs, such as EARTHSCOPE.
GP14A-05 17:15h
An Experimental Study of the Transient Electromagnetic Response of Connected Conductor Networks
The oft-held view of electromagnetic induction as a geophysical technique simply to register bulk electrical conductivity of the ground warrants further attention. The role played by ground inductance L, for example, is seldom fully appreciated. The time-domain electromagnetic response of an ordinary loop of wire demonstrates the important role of inductance. A complete buried loop can generate a sizable electromagnetic response, as any successful beachcomber knows. However, a through-going cut made in the loop does not remove any conducting material from the ground but severely affects the inductance of the loop and renders negligible its electromagnetic response. These ideas can be extended to interconnected conductor networks which are idealizations of fractured rock formations, heterogeneous aquifers, or partial melt in the upper mantle. A field study was performed using several metal wire and ring networks. Network percolation has little impact on the bulk transient electromagnetic response. Instead, significant factors are the total area of closed loops coupled to the primary flux and mutual inductance between loops. A so-called "critical" cut made in a wire mesh network to render it non-percolating actually has a negligible effect on the bulk response if it does not cut through any major loops contained within the network. There may be huge implications of these findings for magnetotelluric data interpretation.
GP14A-06 17:30h
Deconvolution Of TEM Signals Using A Truncated SVD Approach
In transient electromagnetics (TEM) the influences of the measurement equipment and the transmitter current function on the measured signal are usually subsumed as system response. To interpret TEM data by means of inversion the system response is convolved with the synthetic transient before the deviance between synthetic and measured transients are calculated. This approach is preferred to deconvolution because of its numerical stability. To perform the mentioned convolution synthetic values must be calculated for times at least one decade prior to the times measured in the field. Whereas this is not a problem in 1D-inversions it becomes a problem in 2D- or 3D-inversions as the additional decade in time either slows down the forward calculations or even may cause instabilities in the calculation process. To avoid this problems the field data sets have to be deconvolved. As typical deconvolution techniques are proven to be unstable with TEM data we developed a stable technique based on the fit of a sum of exponential functions to the transients. The parameters of the fit build an over-determined linear system that is solved using a truncated SVD approach.
GP14A-07 17:45h
LOTEM and SHOTEM measurements at the Dead-Sea-Transform within the DESERT-project
Within the Dead-Sea-Rift-Transec-(DESERT)-project seismic, seismological, electromagnetic, gravity, magnetic, geodynamic and geological studies were done at the Dead Sea Rift/Dead Sea Transform (DST) to give answers to the questions: "How do shear zones work and what controls them?". Among large scale insights of the DST structure which were derived e.g. from magnetotelluric(MT)- and wide angle seismic-measurements many interesting results on mid and small scales were obtained within the project. Most of the mid and small scale experiments were located on the Araba-fault, which is the main fault of the DST-system between the Dead- and Red-Sea. In this region, we carried out transient-electromagnetic-measurements (TEM) in the years 2002 and 2004. We used two different methods with different penetration depths. With short-offset-TEM (SHOTEM) measurements we explored the conductivity structure of the fault in near surface regions from several meters up to 100 m depth. Using Long-Offset-TEM (LOTEM) measurements we resolved conductivity structures from 100 m up to three kilometers. The field setups of both methods were designed to allow 2-D interpretation. All TEM profiles cross the fault approximately in the middle. The LOTEM setup consisted of a 10 km long profile with four grounded dipole transmitters (two on each side of the fault) and 64 receiver stations. With the SHOTEM method we realized four parallel profiles of 1 km length with station spacing of 50 m station in central-loop configuration. To further enhance the spatial resolution in the central part of each profile, we carried out in-loop measurements with different positions of the receiver. Most of the SHOTEM measurements were realized with a new 3-component-TEM antenna attached to a three channel Nanotem receiver. Using this setup, we were able to record horizontal and vertical signals simultaneously. The design and construction of this new 3-component TEM antenna was done at our institute in Cologne. With 1D inversions of the LOTEM data we obtained models showing similar conductivity features to the MT results. On the west side of the Araba fault our data can confirm the existence of a conductive feature with resistivities below 10 Ohmm in a depth of about 1 km. This structure seems to disappear on the east side of the fault. 1-D results from the SHOTEM measurements across the fault indicate conductivity structures which correlate with velocity structures derived by seismics. Furthermore, our in-loop measurements seem to confirm the existence of a conductive vertical near-surface anomaly at the fault trace. On the southernmost profile, this anomaly is located at the same position where a seismic experiment found guided waves. The existence of guided waves is correlated to the damage zone of the fault. The damage zone itself was expected to be a conductive feature. The final interpretation of the recorded TEM data will be done using 2-D models. For this purpose we develope a new 2.5-D time-domain inversion code. The inversion code was tested with synthetic data and the results are promising. Due to the huge memory demand we currently are running inversions using subsets of the measured data. With a 2.5 inversion of the full data we expect a more detailed insight in the conductivity structure of the DST.