SSsPmp: Can We Expand the Applicable Epicentral Distance (Δ) for Virtual Deep Seismic Sounding?

Date: 4/26/2019

Time: 06:00 PM

Room: Grand Ballroom

Virtual Deep Seismic Sounding (VDSS) infers crustal depth from the travel-time difference (T_VDSS) between teleseismic direct S and SsPmp (Tseng et al., 2009). SsPmp is produced by s-to-P conversion at the free surface forming a virtual source, followed by wide-angle P-to-p reflection at the Moho. Because the Pmp leg undergoes total internal reflection at the Moho, SsPmp has large amplitude and is often directly readable from individual band-pass filtered seismograms. However, to achieve the wide-angle reflection requires a ray parameter similar to the near-Moho 1/V_P, and suitable sources for VDSS are typically deep enough (> c. 60 km) and within a narrow range of Δ (~ 30° to 50°). VDSS has been successfully applied in both western China and across the USA, but many potential target areas on Earth lack suitable subduction seismicity at the appropriate Δ. Yu et al. (2013) used particle-motion analysis to allow VDSS analysis of shallow sources, but their method requires complicated data processing, forsaking the simplicity of VDSS.

Here we attempt to expand the applicable Δ for VDSS sources to c. 65° to 105°, while retaining the inherent simplicity of VDSS, by using SS instead of S phases. We seek to identify SSsPmp following SS, even though SS has a period typically longer (> 20 s) than the usual T_VDSS (< 10 s). Data from GSN station FFC located at Flin Flon, Canada, display SS arrivals that appear to include a delayed compressional component: initial pure shear motions followed closely by a distinct vertical component that oscillates in the same polarity as the radial component. Our recognition of this vertical component as SSsPmp is supported by synthetics for a simple crustal model that produce similar seismograms. The now-popular particle motion analysis of SsPmp is not easily applicable to SSsPmp due to the considerable overlap of shear and compressional motions. Instead, we obtain T_VDSS using synthetics to fit the vertical component of SSsPmp.

Presenting Author: Shiqi Wang

Authors