Real-time event location using model-based estimation of arrival times and back azimuths of seismic phases

Abstract

I have devised a real-time procedure for locating events using an estimation method that analyzes arrival times and back azimuths of phases. The new procedure is applicable to data acquired by local array receivers, such as those used in single-well monitoring as well as by dense receiver networks, and also to noisy waveforms, such as those observed in hydraulic fracturing monitoring if the signal-to-noise ratio is greater than approximately 6 dB. The new procedure uses coalescence microseismic mapping to obtain predictions of arrival times. Based on these predictions, arrival times were estimated by picking the maximum of the ratio of the short-term average to the long-term average of a characteristic function computed for waveforms in an appropriate time window. The estimated arrival times were used in a probabilistic location method, and the probability density function (PDF) of the event location was generated. To locate events for a local array of receivers, the PDFs of event back azimuths obtained using polarizations were combined with the traveltime data to remove directional ambiguities. I have developed this method to generate the PDF of event back azimuths using the average of polarization misfits, which are the differences of the measured and computed polarizations for trial event locations, weighted by the signal-to-noise ratio. Synthetic and field data examples of single-well monitoring of hydraulic fracturing, which required the estimation of event back azimuths in addition to arrival times, were evaluated to determine the effectiveness of the new procedure.