Characterization and utilization of heterogeneous ambient noise field for imaging subsurface structure in the Itoshima Peninsula, Japan

Cross-correlation of ambient noise is an effective approach to extract seismic responses between two stations using seismic interferometry. Since we frequently calculate the cross-correlation of the ambient noise assuming homogeneous distribution of ambient noise sources, heterogeneous distribution of the ambient noise sources would interfere in constructing seismic responses in the calculation of the cross-correlation. In this study, we identified the ambient noise sources recorded in a dense seismic array and utilized the information for better subsurface imaging. The seismic array was composed of 50 stations installed in a 480 m ×350 m area in the Itoshima Peninsula, Japan. By analyzing direction of incoming ambient noise, we found that most of ambient noise was generated by traffic from the nearby street. The traffic noise generated surface waves in lower frequencies (< ~10 Hz) and Pwaves in higher frequencies (> ~20 Hz). We also identified high frequency surface (or air) waves generated by a point source at ~60 Hz. This localized noise could be derived from the renovation work because the location of the source was estimated around the renovation site. We then estimated low- and high-frequency surface wave velocities between each station pair. Although we estimated the surface wave velocities in the limited azimuth between the stations due to the localized noise distribution, we estimated reliable surface wave velocities by considering the noise heterogeneity. High resolution maps of the surface wave velocities were tomographically constructed from the surface wave velocities between the station pairs. Thus, identifying the sources of the ambient noise acquired with dense seismic arrays is effective to improve the ability of ambient noise data to image subsurface structures. It also contributes to the design of seismic arrays in further ambient noise surveys.