Numerical and field experiments for virtual source tomography, Perth Basin, Western Australia

Abstract

A virtual source method (VSM) field experiment was performed at the Mirrabooka Trial Aquifer Storage and Recovery Site in Perth Basin, Western Australia. The experiment used hydrophones deployed simultaneously in two adjacent vertical fibreglass-reinforced plastic monitoring wells. The objective was to provide detailed P-wave velocities between two wells using conventional vertical seismic profiling equipment. It was hoped that the recovery of detailed velocity distribution would provide insight into the distribution of sand and clay above and within a highly heterogeneous injection interval. For the purpose of validating the processing methods used and to gain insight into the radiation pattern of the virtual source, the field experiment was duplicated with finite element numerical modelling. For both numerical and field experiments the seismic energy was propagated using 150 surface source positions with 2 m source point spacing. The seismic energy was recorded simultaneously at two vertical boreholes with 23 hydrophones. The hydrophones on each string were spaced at 10 m intervals. For the numerical model, near-surface velocities were obtained from a refraction seismic survey. All other velocities were derived from acoustic wire-line logging and zero-offset VSP. The thickness of the unsaturated zone in the near-surface layer was approximately 5 m, with P-wave velocities ranging from 60 to 800 m/s. Beyond this was saturated sand/sandstone in which the P-wave velocity was close to 1600 m/s. We directly compare the velocity distributions derived from field and numerical modelling experiments and demonstrate that the virtual source method applied to dual vertical wells has considerable potential. Further analysis with numerical modelling indicates that detail in the crosswell velocity tomogram can potential be pushed to an even higher level of resolution by using dense receiver arrays.