Preprocessing for Reservoir Seismicity Location: Rotokawa Geothermal Field, New Zealand

Abstract

We are investigating bulk seismic properties and fault structure in the Rotokawa geothermal field, New Zealand, by accurately locating and characterizing microseismicity recorded by temporary seismometer arrays, deployed in 2006 and 2008-2009. Analysis approaches used in microearthquake location at Rotokawa and other similar fields need to allow for high background noise levels, complex sub-surface geology, and the high attenuation of near-surface volcanic deposits often experienced in New Zealand. The location problem is exacerbated by high sub-surface temperatures, which currently negate the use of down-hole sensors for any length of time. Our focus in the analysis workflow is on increasing the signal-to-noise ratio of data recorded by surface seismometers, to enable higher accuracy for P- and S-wave arrival picks, dealing with data which occasionally has strongly polarized background noise. We can do so through the use of adaptive polarization and S-transform based filtering before slowness estimation, followed by rotation of 3-component data into ray coordinates (L-, T-, and Q- components) before S-wave picking. Waveform cross-correlation is then carried out on waveform data, before double-difference relocation of the microearthquakes, allowing for 3D heterogeneous velocity structure.