Seismic source mechanism inversion from a linear array of receivers reveals non-double-couple seismic events induced by hydraulic fracturing in sedimentary formation Zuzana Jechumtálová a, ⁎, Leo Eisner b,c a Geophysical Institute, Ac. Sci. of Czech Rep., Boční II 1401, Prague 4, 141 31, Czech Republic b Formerly Schlumberger Cambridge Research, High Cross, Madingley Road, Cambridge, CB3 0EL, United Kingdom c MicroSeismicinc, Inc., 800 Tully Road, Houston, TX 77079, United States abstract article info Article history: Received 14 September 2007 Received in revised form 17 June 2008 Accepted 14 July 2008 Available online 20 July 2008 Keywords: Moment tensor inversion Non-double-couple mechanisms Hydraulic fracturing Induced seismicity We developed a new method to invert source mechanisms of seismic events observed from a single (vertical or near vertical) array of receivers. This method was tested on a synthetic dataset and applied to a seismic dataset acquired during hydraulic fracturing of a sedimentary formation. We show that a crack-opening seismic event recorded on a single vertical array can be inverted as a pure shear seismic event, indicating that the best fitting shear mechanisms inverted from seismic data acquired along a vertical borehole provide misleading information. We found that the majority of the analyzed source mechanisms of the induced events are not consistent with pure shear faulting as previously thought. The non-shear source mechanisms are required even when seismic noise, location uncertainty, and uncertainty in the attenuation structural model are considered. © 2008 Elsevier B.V. All rights reserved. 1. Introduction Hydraulic fracturing is an operation in which a fluid is pumped down a well and into a formation under high enough pressure to cause the formation to open a new fracture. At injection intervals the pressure of the injected fluid usually exceeds the minimum stress thus allowing tensile opening. This procedure is routinely used to increase connectivity and conductivity of the oil and gas bearing reservoirs (usually sedimentary formations, e.g., Phillips et al., 1998) or to create flow paths in geothermal reservoirs (usually volcanic or other non-sedimentary formations, e.g., Pearson, 1981). Furthermore, to keep these fractures conductive solid particles known as proppant are pumped with the fluid into the treated formation. These particles are up to a few millimeters in size, and usually several tens of cubic meters are injected into the formation during a given treatment. The hydraulic fracture treatments induce weak seismic events (commonly known as microearthquakes). However, it is still poorly understood how the injected fluids induce microearthquakes. A key step in resolving this issue is to invert for the source mechanisms of the induced microearthquakes. This can be done by inverting full waveforms or amplitudes of the seismic phases (e.g., Dahm et al., 1999; Jechumtálová and Šílený, 2001) or amplitude ratios of the body-wave phases (e.g., Foulger et al., 2004). The waveform or amplitude inversion is suitable for receivers situated in boreholes or where usually velocity model of the medium is well constrained. However, considering the everlasting uncertainty in knowledge of the medium, source duration, hypocenter location, etc. (which may result in inability to reproduce waveforms), it is more robust to parameterize these waveforms by amplitudes and invert the amplitudes of seismic phases. The moment tensor inversion of amplitude ratios is more suitable for surface stations. It is not only robust to velocity model as long as P- and S-waves travel along the same paths, but also is more sensitive to good receiver coverage of the focal sphere (boreholes provide a very poor focal coverage). The focal coverage is a significant problem because the induced microearthquakes are very weak and observation points are limited (usually to only one observation borehole). Furthermore, the range of induced source mechanisms is much larger than for tectonic earthquakes as the tensile opening may interact with shear faulting. The focal coverage is usually limited to only the closest monitoring borehole as other boreholes are too distant, and signals reaching them are considerably attenuated. Nolen-Hoeksema and Ruff (2001) numerically showed that a single (vertical) array of receivers (i.e., single-azimuth dataset) does not allow for the inversion of the volumetric component of the source mechanism. Therefore, they constrained the inversion to deviatoric mechanisms (i.e., non- volumetric). Vavryčuk (2007) theoretically demonstrated that a single-azimuth dataset cannot resolve the dipole perpendicular to the plane of stations and the hypocenter, i.e. plane of single azimuth. Thus, the deviatoric constraint numerically stabilizes the source mechanism inversion. However, it provides an unphysical result. In this study we exclude the unresolvable component of the source Tectonophysics 460 (2008) 124–133 ⁎ Corresponding author. E-mail address: zs@ig.cas.cz (Z. Jechumtálová). 0040-1951/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.tecto.2008.07.011 Contents lists available at ScienceDirect Tectonophysics journal homepage: www.elsevier.com/locate/tecto