| Abstract |
Roots of extractable geothermal resources (typically less than 3 km) originate at much greater depths, where fluids extract heat from high temperature sources (i.e. magma) and transport it to the surface. The fluid pathways required for this are not well understood; nor is the associated deep-seated permeability. Besides sourcing hydrothermal systems from which electricity can now be generated, these deep-rooted geothermal fluids also represent untapped energy resources themselves, which in the coming years may well be exploited with advances in deep drilling technology. Thus a very good case can be made to expand our knowledge on the roots of geothermal systems to better extract the existing resource base as well as tap into new sources of geothermal energy, including sources beyond current conventional extraction depths. We are investigating the roots of various high temperature geothermal systems present in the Taupo Volcanic Zone (TVZ) in New Zealand’s North Island using magnetotelluric (MT) soundings. Over 23 geothermal systems are known in the TVZ, yet a better understanding of the roots these systems could constrain how they yield an extraordinarily high heat flux. It is currently believed that high temperature convection plumes that extend down to depth of 8 km provide the fluids for these systems. However, much remains uncertain about the basement structure and mechanisms of heat transport and rock permeability below depths of 3 km, which is the present maximum drilled depth. Our ongoing analysis shows clear evidence of deep-seated electrically conductive plumes down to 10 km depth, which are manifested in the near-surface expressions of hydrothermal activity and clear permeable pathways for transporting geothermal fluids. Our experience with the MT data modeling of a complex three-dimensional (3D) environment, such as in the TVZ, showed that it is necessary to analyze the full MT impedance tensor, since the analysis using only principal component tensor data (the off-diagonal components) can produce artifacts in the imaging process. While using only off-diagonal elements might be justified when the imaging grid is aligned with preferred structural trends in the data, for truly 3D environment this may be difficult if not impossible to achieve. Nevertheless, full MT tensor analysis also has its own challenges, especially how one treats data noise for all components in the impedance tensor, where individual entries can vary by several orders of magnitude at a fixed period. In these instances, the accuracy of noise estimates in the data and weights assigned to the components of the impedance tensor are very important. |