Monitoring Spatial and Temporal Changes in Seismic Velocity of the Florida Karst Aquifer System with Phase Cross-Correlation and Wavelet Cross-Spectral Analysis

Abstract:

Empirical Green’s Functions (EGFs) obtained from cross-correlation of ambient seismic noise provide information of the wave speed of the diffuse surface wave field within the Earth’s interior. However, the traditional methods of cross-correlating amplitude binarized seismic signals and linear stacking fail to provide adequate resolution especially when investigating signals greater than those at microseismic frequencies (f > 0.3 Hz) due to rapid signal attenuation. Here, we apply different data processing methods employing cross-correlation of the instantaneous phase of seismic signals based on the analytic signal theory and time-frequency phase weighted stacking to optimize the wave speed retrieval process. Measuring the spatial and temporal variation in seismic surface wave speed provides a better understanding of the forces and geophysical processes that perturb the stress field of the subsurface. Estimation of seismic velocity changes (dv/v) involves accurate measurements of time-lag shifts between EGFs of time periods of interest and a reference time period. We measure dv/v in a karst aquifer system in north-central Florida, utilizing a recently developed method of wavelet cross-spectral analysis (WCSA), which is efficient in measuring the change in dispersive surface wave speed through time even when there is phase difference greater than half a wavelength between two waveforms. Observations of dv/v are compared to changes in groundwater pore pressure, water table elevation, precipitation, and temperature to explore possible relationships between seismic velocity and hydrologic and environmental changes in the shallow subsurface. The WCSA is a promising data processing technique in seismic interferometry for identifying the spatial distribution of temporal changes of the velocity field. This study provides an effective demonstration of a method for continuous monitoring of dv/v from perturbations in the subsurface with a non-invasive technique over short time periods without the need of an active survey.