| Abstract |
Developing a structural model is a key part of exploring and developing a geothermal resource. Along with the impact structure has on a field’s stratigraphy, faults are commonly evoked as permeability conduits and barriers. Subsequently their likely location is a key input into the reservoir conceptual model, well planning activities, reservoir numerical simulations and, in some cases, reservoir pressure distribution and well transient analysis. Developing a structural model for a volcanic-hosted geothermal system is a challenging undertaking because: (1) surface geophysical techniques typically used in other sectors of the resources industry, particularly reflection seismics, are limited by the highly attenuating clay cap and, if present, two-phase zones; (2) volcanic deposits are challenging to accurately identify because textures are difficult to identify in small drill cuttings where hydrothermal overprinting can be extensive and deposits commonly have complicated paleo-topography ; and (3) in contrast with mineral developments in epithermal environments, stratigraphy is often poorly constrained as geothermal developments typically comprise few wells, recovery of geologic material is commonly impoverished and wireline logging is rarely employed to correct/reconstruct well stratigraphy. It follows that, along with offering a simple and defensible fit to available data, structural models developed under these conditions need to articulate the associated uncertainty. The present paper describes the most recent iteration of the Rotokawa structural model. This model comprises three large north-northeast striking structures and a deep, narrow NE-SW orientated depression, here interpreted as a paleovalley. The Rotokawa structural model has been developed to fit the geology below the Wairakei Ignimbrite (around 1400 mVD), as this is the depth interval of greatest interest to the developer. Each fault in the Rotokawa structural model has an associated uncertainty derived by combining factors which account for our confidence in whether the fault is present at all, as well as the degree that the strike and dip magnitudes can be changed. Overall confidence in the present model is increased by its consistency with the regional structural setting, local micro-seismicity (Sewell et al. 2013) and available image log data. |