Analysis and suppression of passive noise in surface microseismic data

Surface microseismic surveys are gaining popularity in monitoring the hydraulic fracturing process. The effectiveness of these surveys, however, is strongly dependent on the signal-to-noise ratio of the acquired data. Cultural and industrial noise generated during hydraulic fracturing operations usually dominate the data, thereby decreasing the effectiveness of using these data in identifying and locating microseismic events. Hence, noise suppression is a critical step in surface microseismic monitoring. In this thesis, I focus on two important aspects in using surface-recorded microseismic seismic data: first, I take advantage of the unwanted surface noise to understand the characteristics of these noise and extract information about the propagation medium from the noise; second, I propose effective techniques to suppress the surface noise while preserving the waveforms that contain information about the source of microseisms.;Automated event identification on passive seismic data using only a few receivers is challenging especially when the record lengths span over long durations of time. I introduce an automatic event identification algorithm that is designed specifically for detecting events in passive data acquired with a small number of receivers. I demonstrate that the conventional STA/LTA (Short-term Average/Long-term Average) algorithm is not sufficiently effective in event detection in the common case of low signal-to-noise ratio. With a cross-correlation based method as an extension of the STA/LTA algorithm, even low signal-to-noise events (that were not detectable with conventional STA/LTA) were revealed.;Surface microseismic data contains surface-waves (generated primarily from hydraulic fracturing activities) and body-waves in the form of microseismic events. It is challenging to analyze the surface-waves on the recorded data directly because of the randomness of their source and their unknown source signatures. I use seismic interferometry to extract the surface-wave arrivals. Interferometry is a powerful tool to extract waves (including body-wave and surface-waves) that propagate from any receiver in the array (called a pseudo source) to the other receivers across the array. Since most of the noise sources in surface microseismic data lie on the surface, seismic interferometry yields pseudo source gathers dominated by surface-wave energy. The dispersive characteristics of these surface-waves are important properties that can be used to extract information necessary for suppressing these waves. I demonstrate the application of interferometry to surface passive data recorded during the hydraulic fracturing operation of a tight gas reservoir and extract the dispersion properties of surface-waves corresponding to a pseudo-shot gather. Comparison of the dispersion characteristics of the surface waves from the pseudo-shot gather with that of an active shot-gather shows interesting similarities and differences. The dispersion character (e.g. velocity change with frequency) of the fundamental mode was observed to have the same behavior for both the active and passive data. However, for the higher mode surface-waves, the dispersion properties are extracted at different frequency ranges.;Conventional noise suppression techniques in passive data are mostly stacking-based that rely on enforcing the amplitude of the signal by stacking the waveforms at the receivers and are unable to preserve the waveforms at the individual receivers necessary for estimating the microseismic source location and source mechanism. Here, I introduce a technique based on the τ − p transform, that effectively identifies and separates microseismic events from surface-wave noise in the τ −p domain. This technique is superior to conventional stacking-based noise suppression techniques, because it preserves the waveforms at individual receivers. Application of this methodology to microseismic events with isotropic and double-couple source mechanism, show substantial improvement in the signal-to-noise ratio. Imaging of the processed field data also show improved imaging of the hypocenter location of the microseismic source. In the case of double-couple source mechanism, I suggest two approaches for unifying the polarities at the receivers, a cross-correlation approach and a semblance-based prediction approach. The semblance-based approach is more effective at unifying the polarities, especially for low signal-to-noise ratio data.