| Title | 3D Inversion of MT Data in Geothermal Exploration: a Workflow and Application to Hengill, Iceland |
|---|---|
| Authors | Gudni Karl ROSENKJAER and Douglas W. OLDENBURG |
| Year | 2012 |
| Conference | Stanford Geothermal Workshop |
| Keywords | Hengill, geophysics, magnetotellurics |
| Abstract | The magnetotellurics (MT) method is widely used in geothermal exploration to estimate the subsurface resistivity structure. The MT method has proven its ability to map the low resistivity layer of hydro altered smectite clay minerals, whose lower boundary correlates with temperatures of approximately 240°C. This information is used for assessing the resource capacity and targeting favorable drilling sites. In order to recover a subsurface resistivity structure, inversion of MT data is performed to find a resistivity model that can explain the data within a desired degree of accuracy. In recent years, improvements of multidimensional inversion algorithms, and increased computational power, have made it feasible to routinely perform 3D inversion of MT data during geothermal exploration. This allows more accurate representation of geological structures and an increased confidence in the resulting resistivity model. The inversion process is not “turn-key” and extraction of a best result requires a workflow. In this paper we develop a strategy to invert MT data for geothermal prospects. The specifics of our workflow are motivated by our need to invert MT field data from the Hengill geothermal area in Iceland. Although there are many components in the workflow, one item that requires special investigation is how best to deal with the different components of the MT data tensor. The off-diagonal components are affected by 1D structures as well as multidimensional information and contain sufficient information to obtain reasonable resistivity model. The diagonal components are only related to multidimensional structures and are commonly much smaller in magnitude. We investigate this and other issues as we step through the inversion of the Hengill data set. |