3-D Resistivity Inversion Models Of MT Array Data In The Taupo Volcanic Zone, New Zealand

Abstract

Independent 3-D resistivity inversion models of magnetotelluric (MT) array data located in the Taupo Volcanic Zone (TVZ), New Zealand, have been generated using WSINV3DMT (Siripunvaraporn et al., 2005) and a preconditioned non-linear conjugate gradient algorithm EMGeo (Newman and Alumbaugh, 2000) implemented on a multi-thousand core machine at Lawrence Berkeley National Laboratory. Previously, 2-D and 3-D inversion models of subsets of the total MT array data have imaged the resistivity structure of the upper crust beneath the TVZ, and have resolved quasi-vertical low-resistivity plumes that connect shallow geothermal fields to an inferred magmatic heat source that lies below the brittle-ductile transition at ~6-8 km depth (Bertrand et al., 2012; 2013). The TVZ in the central North Island of New Zealand is the largest rhyolitic magmatic system on Earth and hosts more than 20 high-temperature liquid-dominated geothermal systems (Bibby et al., 1995; Wilson et al., 1995). Here, we focus on assessing two independent 3-D resistivity inversion models generated using the WSINV3DMT and EMGeo algorithms, each implemented with different control parameters (e.g. error-floors, dataweights, model mesh etc.). We use 3-D inversion models from each code to compute synthetic MT phase tensors for comparison with the measured values. By comparing the synthetic and measured phase tensors, we can efficiently visualize the model fit at all sites in the array, at each period. In this way we can assess the robustness of resistivity structures contained in these MT inversion models and gain insight as to how the different inversion implementations affect the model results.