Seismic Velocity Variations Associated with Injection Stimulation Based on Seismic Ambient Noise Interferometry at the Pohang, Korea, EGS Site

Abstract

The November 15, 2017 Mw5.5 earthquake occurred near the Pohang, Korea EGS site makes it the largest known induced event at an EGS site (Kim et al, Science 2018). Within 10 km from the Pohang EGS site there is a permanent seismic station (PHA2) operating continuously before the induced earthquake. Using the 3-component continuous seismic record from this station we assessed the relative variation of the seismic velocity based on ambient noise interferometry. Auto-correlation and cross-component cross-correlation of the 3-component records generate 6 independent time series (EE, NN, ZZ, EN, NZ and ZE) using a 20-sec time window. These interferometric results can be regarded as the pulse-echo situation, i.e., the source and receiver are co-located at this seismic station (e.g., Claerbout, Geophysics 1968) to ping the seismogenic area in the vicinity of the Pohang EGS site. Using the stretching method (e.g., Lobkis & Weaver, PRL 2003) we can convert the relative travel time change dt/t to the relative velocity change dv/v. A period of 326 days (9/11/2016-8/4/2017) continuous records at PHA2 were used to compute the relative velocity change dv/v. In this period of time two injection stimulations (G1-1, and G2-2, see Woo et al, JGR 2019) were conducted at Pohang EGS. The relative velocity dv/v displays 0.1% variation for both stimulations, a drop prior to the injection and the accompanying seismicity, and recover to a slight rise after the injection. The observed 0.1% velocity change is well above the detection threshold of 10-4, as validated by the velocity response to the tidal strain (Sens-Schönfelder & Eulenfeld, PRL 2019). This work demonstrates that ambient noise observation and interferometry provides a cost-effective means for continuous monitoring EGS safety and potential of induced seismicity at EGS sites.