| Description |
Geoelectrical resistivity distributions can be determined by the inversion of transfer functions derived from transient electric and magnetic fields measured at the Earth's surface by an array of receivers. Two of these functions are the magnetotelluric (MT) impedance and the magnetovariational (MV) tipper. Traditional MT inversions can be distorted by near-surface inhomogeneities (NSI) and are not sensitive to deep conductors; however, MV inversions are both sensitive to deep conductors and far less susceptible to distortion due to NSI. In this thesis, MT impedance and MV tipper data are jointly inverted in three dimensions (3D) to overcome this problem of NSI. A synthetic model study and three examples of varying size and scope are presented. Results from four combinations of transfer functions are compared for each example: principal impedance, principal impedance plus tipper, full impedance, and full impedance plus tipper. The integral equation (IE) method is used in forward modeling and the re-weighted regularized conjugate gradient method (RRCGM) is used in the inversion. We demonstrate that the inclusion of the tipper data in the joint 3D inversion of impedance can provide more accurate information about the location, size, and shape of deep geoelectrical anomalies. The synthetic model study validates the feasibility and performance of the joint inversion methodology. The first example analyzes an audio-MT data set collected in the MacArthur River Valley of Saskatchewan, Canada. The targets in this case are a graphitic fault zone and several known deposits of uranium in the form of uraninite. The geology of this area is well understood. The second example analyzes MT data from the Earthscope project gathered in southern Alberta, Canada. The targets in this example are a large conductive anomaly underlying the Archean Loverna Block and a zone of discreet conductors coincident with the Red Deer High-an upper crustal magnetic anomaly. The geology of this area is not as well understood, and several geophysical studies are referenced. The third example analyzes MT data from the Earthscope project on a much larger scale. Our inversion domain in this example spans the northwestern United States and includes the data from southern Alberta, Canada from the second example. Targets in this example include a subduction zone and backarc, a relic slab curtain, resistive cratons and their intermediary suture zones, and the Yellowstone hotspot. In all cases, the targets are recovered; however, the joint inversion of full impedance and tipper data appears to provide the most accurate information about the size and shape of deep conductors. iv |
