This file was prepared 2020/08/06 by Sara Beth L. Cebry (sbl84@cornell.edu) This file supplements data associated with the publication: "Seismic swarms produced by rapid fluid injection into a low permeability laboratory fault" Earth and Planetary Science Letters Authors: Sara Beth L. Cebry (sbl84@cornell.edu) and Gregory C. McLaskey (gcm8@cornell.edu) -------------------------- Dataset Description: These data are from Laboratory Earthquake Experiments from the Cornell 0.76 m apparatus in support of the following research: Fluid injection, from activities such wastewater disposal, hydraulic stimulation, or enhanced geothermal systems, decreases effective normal stress on faults and promotes slip. Nucleation models suggest the slip at low effective normal stress will be stable and aseismic—contrary to observed increases in seismicity that are often attributed to fluid injection. We conducted laboratory experiments using a biaxial loading apparatus that demonstrate how an increase in fluid pressure can induce “stick-slip” events along a preexisting saw-cut fault in a poly(methyl methacrylate) (PMMA) sample. We compared slip events generated by externally squeezing the sample (shear-triggered) to those due to direct fluid injection (fluid-triggered) and studied the effects of injection rate and stress levels. Shear-triggered slip events began on a localized nucleation patch and slip smoothly accelerated from slow and aseismic to fast and seismic. Fluid-triggered slip events initiated far more abruptly and were associated with swarms of tiny foreshocks. These foreshocks were able to bypass the nucleation process and jump-start a mainshock resulting in an abrupt initiation. Analysis of these foreshocks indicates that the injection of fluid into a low permeability fault promotes heterogeneous stress and strength which can cause many events to initiate—some of which grow large. We conclude that while a reduction in effective normal stress stabilizes fault slip, rapid fluid injection into a low permeability fault increases multi-scale stress/strength heterogeneities which can initiate small seismic events that have the potential to rapidly grow, even into low stress regions. -------------------------- When utilizing this data, please cite as listed below, and provide reference to one or more of the following associated publications: Dataset: Cebry, S. B. L., and McLaskey, G. C. (2020) Data from: Seismic swarms produced by rapid fluid injection into a low permeability laboratory fault [dataset]. Cornell University eCommons Repository. https://doi.org/10.7298/enph-mx14 Publications: Cebry, S. B. L., and McLaskey G. C. (2020) Seismic swarms produced by rapid fluid injection into a low permeability laboratory fault, Earth and Planetary Science Letters, accepted. -------------------------- This work was sponsored by National Science Foundation grant EAR-1847139. -------------------------- These data are shared under a Creative Commons Universal Public Domain Dedication (CC0 1.0); the data will be openly available for re-use, modification and distribution; proper attribution to the original data creators is expected. See citation information above. -------------------------- File labelling: Experiments were conducted in runs which consist of series of events. For catalog purposes a run called “8MPa2_5mLm” denotes the second sequence run at ~8 MPa normal stress with water injection at 5 mL/min. Files contain mechanical and slip sensor data unless they are labeled with "PZ" at the end of the file name (for example "1MPa_5mLm_PZ") in which case they contain piezoelectric sensor data. Piezoelectric data was recoreded in trigger mode and 8MPa_5mLm data were quite large and were saved as four seperate files. "8MPa2_5mLm_PZ1" and "8MPa2_5mLm_PZ2" contain continuous data and shear-triggered events, while "8MPa2_5mLm_PZ3" and "8MPa2_5mLm_PZ4" contain fluid triggered events. Piezoelectric data was also recorded for a 3/32" diameter steel ball being dropped on the top surface of the samples. This data was recorded as a single trigger and saved as "3p32inball". It does not have any corresponding mechanical or slip sensor data. Slip data: Each .mat file contains a MATLAB structure with 2 fields: ‘t’ and ‘data’. ‘t’ is a time vector describing the time stamp of each data point found in ‘data’. All data were recorded at 50 kHz and then averaged down to 5 kHz, which is provided here. ‘data’: Rows 1-8 are data recorded from eddy current sensor channels E1 - E8, respectively. E1 is near the forcing end (North) and E8 is near the leading edge (South) of the sample. All data is in units of Volts and the conversion factor is 128 microns/V. ‘data’: Rows 9 - 10 are data from hydraulic pressure sensors in the array of shear loading cylinders (North side) and normal loading cylinders (East side), respectively. Data is in Volts where 5 V is 10,000 psi. The conversion from Voltage to sample average stress is 9.3 MPa/V for the normal stress (East-side cylinders) and 4.7 MPa/V for the shear stress (North-side cylinders). ‘data’: Rows 11 - 13 are data from pressure sensors connected to the injection well, north side monitoring well, and south side monitoring well, respectively. Injection well data is in Volts where 5 V is 10,000 psi, resulting in a conversion factor of 13.8 MPa/V. Monitoring well data is in Volts where 5 V is 3,000 psi, resulting in a conversion factor of 4.1 MPa/V. ‘data’: Row 14 was used for a piezoelectric sensor signal, PZ3, that was split and recorded for precise time synchronization between this data and the piezoelectric sensor data that was recorded on a separate digitizer. Sensor Setup: Below are the approximate coordinates of the eddy displacement sensors E1-E8. The origin of the coordinate system is the top NE corner of the stationary block. Sensor X (m) Y (m) Z (m) E1 0.03 0 0 E2 0.13 0 0 E3 0.23 0 0 E4 0.33 0 0 E5 0.43 0 0 E6 0.53 0 0 E7 0.63 0 0 E8 0.73 0 0 Piezoelectric sensor data: Each .mat file is a MATLAB structure with 3 fields: ‘t’, ‘triggerTime’ and ‘data’. ‘t’ is a time vector describing the time stamp of each data point found in ‘data’. ‘triggerTime’ indicates the time the trigger started relative to the start of the experiment. Continuous data has a ‘triggerTime’ of 0 s. Ball drop data does not have any corresponding continuous data and therefore has a 'triggerTime' of 0 s. Continuous data is recorded at 50 kHz then averaged down to 5kHz while saving. Trigger blocks are recorded in trigger mode at 5 MHz. ‘data’: Rows 1-8 are the output from sensors PZ1-PZ8 recorded with a digitization range of -100 mV to 100 mV. ‘data’: Rows 9-16 are the output from sensors PZ1-PZ8 recorded with a digitization range of -10 V to 10 V. PZ sensor locations: Below are the coordinates of the piezoelectric sensors PZ1-PZ8. The origin of the coordinate system is the top NE corner of the stationary block. Sensor X (m) Y(m) Z(m) PZ1 0.325 0.3 0 PZ2 0.375 0.3 0 PZ3 0.425 0.3 0 PZ4 0.475 0.3 0 PZ5 0.300 0.3 -0.076 PZ6 0.350 0.3 -0.076 PZ7 0.400 0.3 -0.076 PZ8 0.450 0.3 -0.076