← Back to Sessions

Investigating the Influence of Site Effects and Spatial Stacking on Source Parameter Estimation for Induced Earthquakes Using a Large-N Array

Session: Mechanisms of Induced Seismicity: Pressure Diffusion, Elastic Stressing and Aseismic Slip II

Type: Oral

Date: 4/19/2021

Presentation Time: 05:00 PM Pacific

Description: 

We take advantage of the high station density of the LArge-n Seismic Survey in Oklahoma (LASSO) to investigate how source parameter estimation for induced earthquakes would be affected by: (1) local site responses, and (2) spatial stacking. With ~1825 stations in a 25 km x 32 km region, the LASSO array recorded more than 1500 local events with unprecedented resolution in an area of extensive wastewater injection during spring 2016. These data are ideal for analyzing how spatially varying conditions across the array affect our estimation and if we can improve it by stacking.

We find that the earthquake corner frequencies (fc) estimated using the single spectrum method at each station negatively correlate with local site amplifications. Site amplifications are estimated at each station using 5 second RMS amplitude of P coda waves of a regional earthquake (Mw 3.7) 130 km away. The average RMS amplitude deviation among sites is about 50% of the fc deviation for one of the best recorded events. We find that sites with high amplification are typically located on young alluvial sedimentary deposits. Of the parameters estimated using the single spectrum method, corner frequencies are more affected by site amplification, whereas the seismic moments seem to be affected by the radiation pattern. We also apply an Empirical Green's Function (EGF) method to estimate source parameters from spectral ratios and remove the influences of site amplification on fc. We group the stations based on azimuth and distance and stack their spectral ratios and source time functions (STF) to estimate source parameters for some of the best recorded events (Mw 2.2~2.7) within the array. The STFs show gradual spatial variations that describe the rupture directivity, while the fc estimated using the EGF method are likely biased due to high frequency attenuation. We vary the station density, azimuth, and distance range of the sub-groups and analyze the effects on the fitting results.

Presenting Author: Hilary Chang

Student Presenter: Yes

Authors