The ability to detect small microearthquakes and identify their P and S phase arrivals is a key issue in hydrofracture downhole monitoring because of the low signal-to-noise ratios.We apply an array-based waveform correlation approach (matched filter) to improve the detectability of small magnitude events with mechanisms and locations similar to a nearby master event. After detecting the weak events, we use a transformed spectrogram method to identify the phase arrivals. We have tested the technique on a downhole monitoring dataset of the microseismic events induced by hydraulic fracturing. We show that, for this case, one event with a signal-to-noise ratio around 6dB, which is barely detectable using an array-stacked short-time average/long-time average (STA/LTA) detector under a reasonable false alarm rate, is readily detected on the array-stacked correlation traces. The transformed spectrogram analysis of the detected events improves P and S phase picking.

A borehole network consisting of 5 monitoring wells was used to monitor the induced seismicity at a producing petroleum field for a period of about 11 months. Nearly 5400 microseismic events were analyzed and utilized in imaging the reservoir based on a new doubledifference (DD) seismic tomography. The DD tomography method simultaneously solves for event locations and Vp, Vs, and Vp/Vs models using absolute and differential P, S and S-P arrival times. Microseismicity in the field was primarily caused by compaction of the reservoir in and above the gas bearing formation and was distributed along the two major northeastsouthwest (NE-SW) faults in the field. The model resolution analysis based on the checkerboard test and the resolution matrix showed that the central part of the model was relatively well resolved for the depth range of 0.7 to 1.1 km. Clear velocity contrasts were imaged across most parts of the two NE-SW faults. Vp/Vs ratio estimations from the tomographic inversion were low (<1.75) in the shallow depth range, likely due to lithology and gas content, whereas they were large (>1.75) in the deeper part of the model, likely due to fluid saturated formation. In this study seismic tomography showed a great potential for reservoir imaging and property estimation using induced seismicity.2