Seismic exploration-scale velocities and structure from ambient seismic noise (> 1 Hz)

The successful surface waves retrieval in solid-Earth seismology using long-time correlations and subsequent tomographic images of the crust have sparked interest in extraction of subsurface information from noise in the exploration seismology. Subsurface information in exploration seismology is usually derived from body-wave reflections >1Hz, which is challenging for utilization of ambient noise. We use 11h of noise recorded in the Sirte basin, Libya. First, we study the characteristics of the noise. We show that the bulk of the noise is composed of surface waves at frequencies below 6Hz. Some noise panels contain nearly vertically traveling events. We further characterize these events using a beamforming algorithm. From the beamforming, we conclude that these events represent body-wave arrivals with a fairly rich azimuthal distribution. Having body-wave arrivals in the noise is a prerequisite for body-wave reflections retrieval. We crosscorrelate and sum the recorded ambient-noise panels to retrieve common-source gathers, following two approachesusing all the noise and using only noise panels containing body-wave arrivals likely to contribute to the reflections retrieval. Comparing the retrieved gathers with active seismic data, we show that the two-way traveltimes at short offsets of several retrieved events coincide with those of reflections in the active data and thus correspond to apexes of reflections. We then compare retrieved stacked sections of the subsurface from both approaches with the active-data stacked section and show that the reflectors are consistent along a line. The results from the second approach exhibit the reflectors better.