Stanford Geothermal WorkshopFebruary 8-10, 2027

The magnetotelluric (MT) method is widely recognized as one of the key exploration tools for investigating geothermal reservoirs. With the development over the past decade of three-dimensional (3D) inversion codes for MT data that employ the finite-element method (FEM) for forward modeling, increasingly accurate resistivity models can now be obtained. Particularly, FEM inversion codes can incorporate topography into 3D meshes without losing numerical accuracy significantly. However, applying 3D FEM inversion codes to field MT data is not straightforward because various parameters, including mesh settings, must be appropriately defined. The resulting resistivity models may depend on these parameter settings. In this study, we investigated the effects of different mesh settings, including tetrahedral versus hexahedral elements and different element sizes, on 3D inversion results for MT data acquired in the Ogiri geothermal area. The dataset consisted of MT measurements at 244 stations. We used all components of the impedance tensor and tipper data at 16 frequencies ranging from 0.00293 to 96 Hz. The 3D FEM inversion code FEMTIC was used for the analysis. FEMTIC supports both tetrahedral (Tetra) and distorted non-conforming hexahedral (DHexa) elements for mesh construction. We prepared two meshes with different element sizes for each element type. For the Tetra meshes, the smallest elements around the MT sites were less than 25 m and 40 m, respectively. For the DHexa meshes, the smallest horizontal element sizes around the MT sites were 31.25 m and 62.5 m, respectively, while the vertical element size near the ground surface was 25 m. All other inversion settings, including the input data, error floors, and treatment of galvanic distortion, were identical. Comparison of the final resistivity models revealed significant differences between the Tetra and DHexa meshes in features important for geothermal interpretation, particularly the geometry of the shallow low-resistivity cap rock and resistivity anomalies within the deeper high-resistivity reservoir zones. Differences in element size within both the Tetra and DHexa meshes also produced meaningful variations in these resistivity structures. Although the causes of these differences could not be conclusively identified, our results demonstrate that the choice of mesh type and element size can significantly affect the resulting resistivity models. Therefore, the sensitivity of inversion results to mesh setting should be carefully evaluated when interpreting the resistivity structure of geothermal reservoirs.

Topic: Geophysics