FAULT-CONSTRAINED JOINT SEISMICITY LOCATION AND VELOCITY TOMOGRAPHY

Estimation of seismicity location is critical to reveal the spatiotemporal characteristics of geological faults. There have been well-established hypocenter location methods based on traveltime or full-waveform backpropagation; some methods can also jointly estimate a velocity model during the seismicity location inversion process. However, existing methods in general do not account for the physics or geometry of faults and the interpretation of fracture/faults are usually performed posterior to the location procedure. Hence, estimated spatiotemporal locations of different seismicity events can be arbitrary and do not explicitly or implicitly correlate with one another. We develop a novel method for estimating seismicity location by integrating fault geometry constraints iteratively in the least-squares inversion. Specifically, we cluster estimated seismicity locations automatically, fit each of the clustered events to a meaningful curved surface, and use the iteratively updated fault surfaces to constrain the inverted seismicity locations, in addition to minimizing the traveltime absolute misfit or double-difference misfit. We utilize a factorized eikonal equation to compute the forward and adjoint traveltime fields to pertain higher accuracy and flexibility in complex geological models. Our validation results show that by imposing fault-surface constraint derived from automatically clustered seismicity, the seismic location estimation with or without joint velocity tomography can be improved and have a better interpretability in terms of fault geometry-seismicity correlation.