An Optimized Workflow for Source Localization and Joint Velocity Inversion using Wavefront Attributes

Abstract:

The localization of passive seismic sources plays an essential role in seismology and exploration. Many current localization techniques rely on prior knowledge of the underlying velocity model which, however, is not often available in practice. We present a new workflow based on wavefront attributes, which locates passive seismic sources and build the velocity model simultaneously. The localization/tomographic scheme only requires surface velocity information for estimating wavefront attributes which locally describe the kinematics of the incident wavefield of a seismic event.

The workflow contains two primary steps. The first step is a coherence analysis to evaluate wavefront attributes comprising slopes and curvatures as well as source excitation times. These quantities serve as the input for the tomography in the second step. This step includes refining the previously estimated source excitation times of the recorded events using particle swarm optimization, enforcing source focusing by time-reversal techniques, and estimating the velocity distribution in a simultaneous fashion.

Numerical case studies for several complex models, such as SEG-salt model and sigsbee 2A model reveal the functionality of the approach. The refined source excitation time errors are within the prevailing period of the source signal and the maximum location errors are within the dominant wavelength of the signal, i.e. within the resolution limits for band limited signals. The maximum relative velocity errors are about 10%. The proposed scheme for simultaneous location and velocity model building is applicable from the micro-seismic scale to seismological problems.