GP32A-01 INVITED
EMSCOPE Electromagnetic Component of EarthScope Backbone and Transportable Array Experiments 2006-2008
USArray (http://www.iris.edu/USArray) in conjunction with EMSOC (Electromagnetic Studies of the Continents) (http://emsoc.ucr.edu/emsoc) is installing magnetotelluric (MT) stations as part of Earthscope. The MT component of Earthscope consists of permanent (Backbone) and transportable long period stations to record naturally occurring, time varying electric and magnetic fields to produce a regional lithospheric/asthensospheric electrical conductivity map of the United States. The recent arrival of 28 long period MT instruments allows for the final installation of the Backbone stations throughout the US and yearly transportable array studies. The Backbone MT survey consists of 7 stations spaced throughout the continental US with preliminary installation at Soap Creek, Oregon; Parkfield, California; Braden, Missouri and Socorro, New Mexico.Siting and permitting are underway or completed at stations in eastern Montana, northern Wisconsin and Virginia. These stations will be recording for at least five years to determine electrical conductivities at depths that extend into the mantle transition zone. The first transportable array experiment was performed in the summer and fall of 2006 in central and eastern Oregon (Oregon Pilot Project) using equipment loaned from EMSOC. Thirty-one long period MT stations were recorded with 14 to 21 day occupations. Preliminary 3D inverse models indicate several lithospheric electrical conductivity anomalies including a linear zone marked by low-high conductivity transition along the Klamath-Blue Mountain Lineament associated with a linear trend of gravity minima. High electrical conductivity values occur in the upper crust under the accreted terrains in the Blue Mountains region. The second transportable array experiment was performed in the summer and fall of 2007 and completes coverage of the Oregon, Washington, and western Idaho, targeting the Cascadia subduction zone, Precambrian boundaries, and sub-basalt lithologies. The 2008 transportable MT experiment will focus on the Snake River Plain and the Yellowstone Hot Spot. The disposition of future USArray magnetotelluric geotransects will be the subject of an upcoming NSF-supported planning workshop. Time series are available now from the IRIS data center (www.iris.edu/data), and magnetotelluric transfer functions will soon be available.
GP32A-02
Three-Dimensional Forward Modeling of Magnetotelluric Data Over Cratonic Lithosphere and Attendent Geological Structures: Case Study of the Zimbabwe Craton
The Southern African MagnetoTelluric Experiment (SAMTEX) is covering a huge area - containing parts of South Africa, Namibia and nearly the whole of Botswana - with acquisition of magnetotelluric (MT) data. The project's aim is to gain more information on the lithospheric geometries of the geological structures of this region which contains some of the oldest lithospheric pieces on Earth - the cratons (e.g. Kaapvaal and Zimbabwe cratons). Since the standard modeling of MT data is only in two-dimensions and the recently developed three-dimensional inversion programs require a lot of computation time and high speed computers with large memories, 3D forward modeling is a good compromise on the way to full 3D interpretation of MT data. Forward modeling of subsurface structures similar to the cratonic areas of study will give some indication of how the MT responses should look like and if they change significantly when the cratons have a different shape or extent or resistivity. The determined synthetic data then can be compared with the real data collected in the SAMTEX project. The area for the case study is the western edge of the Zimbabwe craton in eastern Botswana. The craton is surrounded by the Damara Mobile Belt and the Magondi Mobile Belt to the north and west and the Limpopo belt to the south. The giant northern Botswana dyke swarm is cross cutting the craton in about WNW to ESE direction. Based on this geological information, a 3D model was created to calculate synthetic MT responses using the forward modeling routine implemented in the 3D inversion program MT3Dinv (developed by the Geophysical Inversion Facility, University of British Columbia). We will show the results of this modeling exercise and compare them with the observations.
GP32A-03
Analysis and 3-D inversion of EarthScope magnetotelluric data : some preliminary results
As part of the EarthScope project in 2006-07 approximately 100 long period (5-15000 s) magnetotelluric (MT) sites were occupied covering Oregon and Washington. In contrast to traditional MT surveys where sites are concentrated along one or a few profiles, in the EarthScope project sites have been occupied on a quasi-uniform 75 km grid. Both the site distribution, and the wide station spacing covering a wide range of geologic environments such as subduction zone at the coast, cascade volcanoes, columbia plateau, high desert into the basin and range demand three-dimensional (3D) inversion and interpretation. Here we present preliminary efforts in this direction. The MT data, which were collected in a series of overlapping arrays, were processed using robust remote reference and multi-station codes to derive MT impedances, along with vertical and interstation transfer functions. We present maps of these transfer functions (i.e., hypothetical events, induction vectors), along with results from distortion, dimensionality, and phase tensor analysis. We also present results from initial 3D inversions of all the four measured tensor components, carried out using the data-space Occam inversion code of Siripunvaraporn et al. (2004). This inversion poses a challenge in terms of the substantial computational resources required for inversion of a data set of this size. One strategy we are trying, and will discuss, is to focus on inversion of data from a reduced number of frequencies, chosen from studying the frequency dependence of induction vectors. Higher error floor was assigned to the static shifted apparent resistivities compared to their phases. This helps in downweighting the apparent resistivities with respect to the phases which in turn reduce the influence of static shifts. Preliminary interpretations based on the inversion results, together with the hypothetical event, induction vector, and phase tensor maps, and results from previous experiments in this area will be presented.
GP32A-04
Constraining the Composition and Thermal State of the Moon from an Inversion of Electromagnetic Lunar Day-Side Transfer Functions
We present a general method to constrain planetary composition and thermal state from a stochastic inversion of long-period electromagnetic sounding data. As an example of our approach, we reexamine the problem of inverting lunar day-side transfer functions to constrain the internal structure of the Moon. We go beyond the conventional approach by inverting directly for chemical composition and thermal state, using the model system CaO-FeO-MgO-Al2O3-SiO2, rather than subsurface electrical conductivity structure, which is only an indirect means of estimating the former parameters. Using Gibbs free energy minimisation, we calculate the stable mineral phases, their modes and densities. The mineral modes are combined with laboratory electrical conductivity measurements to estimate the bulk lunar electrical conductivity structure from which transfer functions are calculated. To further constrain the radial density profile in the inversion we also consider lunar mass and moment of inertia. The joint inversion of electromagnetic sounding and gravity data for lunar composition and selenotherm as posited here is found to be feasible. Our results are not only in line with earlier investigations of the lunar electrical conductivity structure, but also with investigations of the elastic properties (seismic velocities) of the Moon. Our results indicate a bulk composition that is different from the Earth's upper mantle, in line with the giant impact hypothesis for the formation of the Moon. In order to improve future inferences about lunar composition and thermal state, more electrical conductivity measurements are needed especially for minerals appropriate to the Moon, such as pyrope and almandine.
GP32A-05
Scripted Finite Element Methods Applied to Global Geomagnetic Induction
Magnetic field observations from CHAMP, Ø rsted and SAC-C and improved techniques for comprehensive geomagnetic field modeling have generated renewed interest in using satellite and observatory data to study global scale electromagnetic induction in Earth's crust and mantle. The primary external source field derives from variations in the magnetospheric ring current, and recent studies show that over-simplified assumptions about its spatial structure lead to biased estimates of the frequency-dependent electromagnetic response functions generally used in inversions for mantle conductivity. The bias takes the form of local time dependence in the C- response estimates and highlights the need for flexible forward modeling tools for the global induction problem to accommodate 3D time-varying structure in both primary and induced fields. We are developing such tools using FlexPDE, a commercially available script-based finite element method (FEM) package for partial differential equations. Our strategy is to model the vector potential \mathbf{A}, where \mathbf{B} = \nabla × \mathbf{A}, thereby ensuring that \mathbf{B} is divergenceless. We compare our results with existing 2D and 3D analytical solutions to validate the solution techniques. We are able to mitigate time and memory constraints by adaptive refinement of a mesh that is initialized for high resolution in areas of interest. Solutions can be developed in either frequency or time domain, but we use time domain modeling because interpretation of satellite data is complicated by the motion of the satellites through magnetic fields that vary both temporally and spatially. We conduct 3D simulations of electromagnetic induction in a spherically symmetric planet with an asymmetric primary external field. This allows us to document expected differences in response estimates for ground versus satellite instruments. Suitable extensions of these tools are anticipated for use in induction studies of any planet or spherical body, allowing arbitrary spatial and temporal behavior in both primary and induced fields.
GP32A-06
The inner structure of La Fossa di Vulcano (Vulcano Island, southern Tyrrhenian Sea, Italy) revealed by high resolution electric resistivity tomography coupled with self-potential, temperature, and soil CO2 gas measurements
La Fossa cone is an active stratovolcano located on Vulcano Island, in the Aeolian Archipelago (southern Italy). Its activity is characterized by explosive phreatic eruptions and phreato-magmatic eruptions producing wet and dry pyroclastic surges, pumice fall deposits and highly viscous lava flows. Nine profiles of 2D high resolution Electrical Resistivity Tomography (ERT) (electrode spacing 20 meters, with a depth of penetration > 200 meters) were performed across this edifice to image its inner structure. In addition, we also measured the self-potential, the flux of CO2, and the temperature along these profiles. These data provide complementary information to interpret the ERT profiles. The ERT profiles allow to identify the main structural boundaries (and their associated fluid circulations) structuring the shallow architecture of the Fossa cone. The hydrothermal system is identified by very low values of the electrical resistivity (< 20 m). Its lateral extension is clearly limited by the crater boundaries, which are relatively resistive (> 400 m). Inside the crater, it is possible to follow the plumbing system of the main fumarolic area at depth. On the flank of the edifice, a thick layer of tuff is also marked by low resistivity values (in the range 1 to 20 m). The ashes and pyroclastic materials ejected during the XIX Century eruptions and covering the flank of the volcano corresponds to relatively resistive materials (several hundreds to several thousands m). Laboratory measurements are performed to determine the streaming coupling coefficient of the main materials forming the edifice. A 2D simulation of the ground water flow is performed over the edifice using the finite element code Comsol Multiphysics 3.3. Forward and inverse modeling of the self- potential data can be used to put constraints on the flux of water in the flanks of the edifice and inside the crater. The result reveals the very high potentiality of these methods for high resolution imaging of the inner structure of an active volcano. http://www.andre-revil.com/
GP32A-07
Modification of three-dimensional magnetotelluric inversion WSINV3DMT to be applied to seafloor
In recent years, a number of seafloor electromagnetic (EM) experiments have been carried out by using Ocean Bottom Electromagnetometers (OBEMs). The density of marine Magnetotelluric (MT) data has been increasing so that imaging electrical conductivity structures under the seafloor in three dimensions is now feasible. WSINV3DMT (Siripunvaraporn et al., 2005) is known as a full three-dimension inversion code for land MT data, and is based on the data space approach. We try to modify this code so as to be able to invert marine MT data, by including bathymetry features and extending locations of observation sites (the output points) allowing on the seafloor as well as the Earthfs surface. As shown in a separate presentation by Utada et al. (2007, in this meeting), we deployed OBEMs in a two-dimensional array in the Philippine Sea. The aim of this code modification is to apply the modified code to the data set from this array experiment to image three-dimensional conductivity distribution in the deep mantle below. The accuracies of the forward modeling part in the modified WSINV3DMT have been verified with one- dimensional resistivity structures. Relative differences between of the MT impedances estimated by the modified WSINV3DMT and those analytically calculated for horizontal layered structures are all below 1%. It was also found that these differences depend on frequency. In this presentation, we will show the accuracy and reproducibility of the modified WSINV3DMT by using synthetic data for three-dimensional resistivity structures.
GP32A-08
Hydrological Investigations of the Santo Domingo Basin, New Mexico Using Electromagnetic Soundings
Transient electromagnetic (TEM) and magnetotelluric (MT) soundings made during the SAGE (Summer of Applied Geophysical Experience) program in 2006-2007 to assess the hydrologic environment in the arid environment of the Santa Domingo Basin, northern New Mexico. Water wells in the upper 600 m allowed the comparison of the field data with synthetic soundings derived from borehole resistivity logs. Conductive, clay-rich layers of less than 20 ohm-m dominate the upper basin fill except where resistive, presumably sand, layers of several hundred ohm-m are present. Only where the thickness of the resistive layers exceeds their depths can they be resolved by the electromagnetic (EM) methods. Consequently, delineation of the highest quality aquifers is only possible using EM if resistive layers are relatively thick. One- and two-dimensional inversion of the TEM and MT data, respectively, allowed some estimates of the salinity and/or porosity of the thick, predominantly sand, resistive zones using Archie's law. Forward and inverse modeling facilitated sensitivity analyses to evaluate which modeled subsurface features were required by the field data and which features were supported by borehole logs. EM imaging of thick, water-saturated resistive zones, where confined and laterally extensive, reveal aquifers valuable both for their potable water and storage capacity.