Abstract:
This thesis describes the study and modelling of the flow of heat and mass in the geothermal systems within the Taupo Volcanic Zone (TVZ). This was undertaken with the aim of gaining insight into the behaviour of the convective processes that form the numerous high-temperature geothermal fields of the TVZ and of the underlying physical structure of the region. An inverse modelling approach was implemented. This was aided by the use of the iTOUGH2 simulator, which provided automatic model calibration and mathematical optimisation techniques to improve the model match to the field data. The available observation data concerned the surface heat outputs of the TVZs geothermal fields, while the parameters of interest were the underlying permeability structure of the TVZ and the magnitudes of the regions deep heat sources. The initial phase of the modelling process consisted of two stages of two-dimensional modelling, which were performed in order to obtain an approximation of, and so some insight into, the TVZs convective processes. The first stage of this proved that a single 2D model could not sufficiently represent the complexity of the heat flows over the entire TVZ and so four sub-regions were defined. A 2D representative slice model was then created for each of these regions. Each of these models had a good match to the observation values specific to their region. A simple permeability structure common to all four models was maintained throughout all four models, thus providing no evidence against the possibility that the spatial distribution of the geothermal fields within the TVZ could be due to the heating of a uniform porous media from below. 2D modelling is not sufficient to represent the 3D heat transport of the TVZ: the completion of a successful 3D model will be a central aspect of the understanding and modelling of the large-scale convective processes taking place within the TVZ. Therefore work was done to extend the 2D slice models into 3D models of each of the four regions. A limited amount of 3D inverse modelling was achieved on one of these four 3D models.