Optimal production of oil from the South Belridge oil field depends on the accurate characterization of hydraulic fractures. This paper presents an extension of the passive imaging work done by Vinegar et al., (1991). A hydrofracture induced in the South Belridge reservoir was imaged using seismic data acquired from three dedicated observation wells. These wells recorded microseismic events produced during the hydrofracturing process. We had to invert for the event locations using only shear wave arrivals because no compressional wave arrivals were detected.

Our initial inversion assuming a homogeneous velocity model yielded unsatisfactory results. The microseismic event locations failed to reveal a clearly defined fracture plane. However, the excitation of tube waves in the treatment well by nearby microseismic events led to the detection of conical waves by the geophones. These conical wave arrivals provided a means by which to estimate a ‘relative’ velocity field.

Choosing a reference velocity and a reference direction toward one observation well, we were able to calculate relative velocities toward the other observations wells and establish a heterogeneous velocity model. Inverting for the locations of the microseismic events with this new model resulted in a much more defined fracture plane. We now hope to develop a time variable heterogeneous velocity model since the conical wave arrivals occur fairly regularly throughout our data.