Description: In January 2011, a sequence of earthquakes occurred in close proximity to a well, which was being hydraulically fractured in south-central Oklahoma. The hydraulic fracturing of the Picket Unit B Well 4–18 occurred from 16 January 2011 18:43 through 22 January 16:54 UTC. This vertical well penetrated into the mature Eola-Robberson oil field. Earthquakes were identified by cross correlating template waveforms from manually identified earthquakes and cross correlating these templates through the entire operation period of the Earthscope USArray Transportable Array (TA) station X34A. This produced a series of 116 earthquakes, which occurred from 17 January 2011 19:06 through 23 January 3:13 UTC with no other similar earthquakes identified at other times prior to or post-hydraulic fracturing. The identified earthquakes range in local magnitude ( M L ) from 0.6 to 2.9, with 16 earthquakes M L  2 or greater and a b -value of 0.98. There is a strong temporal correlation between hydraulic fracturing and earthquakes. This correlation is strengthened because hydraulic fracturing operations ceased for ~2 days due to bad weather, and earthquakes can be observed to cease during this period and resume after hydraulic fracturing had resumed. Earthquakes were relocated using cross-correlated phase arrivals and bootstrap iterations of hypoDD. Locations were well constrained for 86 earthquakes. These earthquake locations clearly delineate a fault which strikes ~166°, subparallel to the mapped minor fault sets in the area, and dips steeply to the west. The earthquakes appear to have occurred at shallow depths from ~2 to 3 km and within ~2.5 km horizontally of the well. The first earthquake occurred ~24 hrs after hydraulic fracturing began at the well. This delay is consistent with the diffusion of pore pressure in the subsurface over a distance of ~2 km. Online Material: Results from bootstrap hypoDD relocations using cross-correlation phase arrivals.

Description: Variability in seismic instrumentation performance plays a fundamental role in our ability to carry out experiments in observational seismology. Many such experiments rely on the assumed performance of various seismic sensors as well as on methods to isolate the sensors from nonseismic noise sources. We look at the repeatability of estimating the self-noise, midband sensitivity, and the relative orientation by comparing three collocated Nanometrics Trillium Compact sensors. To estimate the repeatability, we conduct a total of 15 trials in which one sensor is repeatedly reinstalled, alongside two undisturbed sensors. We find that we are able to estimate the midband sensitivity with an error of no greater than 0.04% with a 99th percentile confidence, assuming a standard normal distribution. We also find that we are able to estimate mean sensor self-noise to within ±5.6 dB with a 99th percentile confidence in the 30–100-s-period band. Finally, we find our relative orientation errors have a mean difference in orientation of 0.0171° from the reference, but our trials have a standard deviation of 0.78°. Electronic Supplement: Table of dates of the trials used as well as Q – Q plots for the statistics collected from the sensor tests.

Description: Long-period (〉100 s period) seismic data can often be dominated by instrumental noise as well as local site noise. When multiple collocated sensors are installed at a single site, it is possible to improve the overall station noise levels by applying stacking methods to their traces. We look at the noise reduction in long-period seismic data by applying the time–frequency phase-weighted stacking method of Schimmel and Gallart (2007) as well as the phase-weighted stacking (PWS) method of Schimmel and Paulssen (1997) to four collocated broadband sensors installed in the quiet Albuquerque Seismological Laboratory underground vault. We show that such stacking methods can improve vertical noise levels by as much as 10 dB over the mean background noise levels at 400 s period, suggesting that greater improvements could be achieved with an array involving multiple sensors. We also apply this method to reduce local incoherent noise on horizontal seismic records of the 2 March 2016 M w  7.8 Sumatra earthquake, where the incoherent noise levels at very long periods are similar in amplitude to the earthquake signal. To maximize the coherency, we apply the PWS method to horizontal data where relative azimuths between collocated sensors are estimated and compared with a simpler linear stack with no azimuthal rotation. Such methods could help reduce noise levels at various seismic stations where multiple high-quality sensors have been deployed. Such small arrays may also provide a solution to improving long-period noise levels at Global Seismographic Network stations.