Imaging Time Dependent Crustal Deformation Using GPS Geodesy And Induced Seismicity, Stress And Optimal Fault Orientations In The North American Mid-Continent

Abstract

Transient deformation has been observed in a number of different types of tectonic environments. These transient deformation signals are often observed using continuous GPS (CGPS) position time-series observations. Examining transient deformation using CGPS time-series is problematic due to the, often, low signal-to-noise ratios and variability in duration of transient motions observed. A technique to estimate a continuous velocity function from noisy CGPS coordinate time-series of is examined. The resolution of this technique is dependent on the signal-to-noise ratio and the duration or frequency content of the transient signal being modeled. Short period signals require greater signal-to-noise ratios for effective resolution of the actual transient signal. The technique presented here is similar to a low-pass filter but with a number of advantages when working with CGPS data. Data gaps do not adversely impact the technique but limit resolution near the gap epochs, if there is some a priori knowledge of the noise contained within the time-series this information can be included in the model, and model parameter uncertainties provide information on the uncertainty of instantaneous velocity through time. A large transient has been observed in the North-American stable continental interior as a significant increase in the number and moment release of earthquakes through time. This increase in the number of earthquakes has been suggested to be largely related changes in oil and gas production activities within the region as triggered or induced seismicity, primarily from fluid injection. One of the first observed cases of triggered earthquakes from hydraulic fracturing where the earthquakes were large enough to be felt by local residents is documented. The multiple strong temporal and spatial correlations between these earthquakes indicate that hydraulic fracturing in a nearby well likely triggered the earthquake sequence. The largest magnitude earthquake in this sequence was a magnitude 2.9 with 16 earthquakes greater than magnitude 2. The earthquakes in this sequence occurred within 2.5 km of the hydraulic fracturing operation and focal depths are similar to the depths of hydraulic fracturing treatment depths. In addition to the documentation of a transient earthquake signal associated with hydraulic fracturing, the observed focal mechanisms throughout Oklahoma are documented. These focal mechanisms were used to examine the maximum horizontal stress orientations and active fault orientations associated with the increased rates of seismicity observed in the region. Generally, active-fault orientations and the stresses are consistent through broad portions of Oklahoma with one exception, the ongoing Jones earthquake sequence in central Oklahoma that started in 2009. In the Jones earthquake sequence a bi-modal distribution of focal mechanisms are observed. One orientation of active faults observed in the Jones earthquake sequence would not be expected to be active in the observed regional stress field. This unfavorably oriented set of faults appear to be pre-existing structures and activity on these structures may suggest that pore-pressure increases in the sub-surface due to fluid injection in the area make it possible for faults that are not optimally oriented within the regional stress-field to reactivate.