High Frequency Ambient Noise Seismic Data Processing and Quality Control Toward Monitoring Changes in Spatiotemporal Seismic Velocities

Empirical Green's functions (EGFs) obtained from utilizing the characteristics of the diffuse surface wave field provide information of the velocity structure within the Earth's interior. Tomographic studies have identified various scales of subsurface structures based on the dispersive nature of surface waves. However, investigation of shallow depth (< 1km) structures that are more heterogeneous than deeper depths are limited by high signal attenuation and multi-scattering of seismic waves, especially with frequencies greater than microseismic bandwidth. We compare the convergence of EGFs and dispersion measurements of high frequencies (0.5 - 20 Hz) from both synthetic data and observed data measured at a karst aquifer system in O'leno and River Rise Preserve State Parks, in north-central FL, obtained by the conventional cross-correlation and linear-stacking (TCC-LS) method including a 1-bit amplitude normalization with a more recently developed phase-cross-correlation and time-frequency phase-weighted stacking (PCC-PWS) method that was updated to reduce the high computational expense. Moreover, we explore quality control factors including signal-to-noise ratio (SNR) and both lower and upper number of wavelengths (LNWL and UNWL) propagating between interstation spacings to find site-specific optimal cutoffs that improve the network average group velocity curve for the TCC-LS method (SNR: 9, LNWL: 2, UNWL: 40) and PCC-PWS method (SNR: 15, LNWL: 3, UNWL: 50). Finally, we investigate the reproducibility of signal extraction based on the number of stacks between the TCC-LS and PCC-PWS methods to provide information in performing reliable spatio-temporal monitoring applications.