Teleseismic travel-time delays in the Las Vegas basin

摘要

A temporary broadband seismic array was deployed in the Las Vegas basin (LVB), home to one of the fastest growing communities in the United States, to investigate structure in this deep (similar to 5 km) sedimentary basin. To constrain basin velocity structure, we measured the differential travel time of teleseismic P waves by waveform cross correlation relative to a station near the basin's edge. The range of the travel-time delays is significant (up to 0.5 see), and the pattern of travel-time delays is independent of the back azimuth of the incoming energy, suggesting that the near-surface structure controls the delay times. Assuming the reported basin geometry of Langenheim et al. (2001), we modeled the average delay times at the basin stations to estimate the average P-wave velocity structure within the basin. The average times can be modeled with relatively fast P-wave velocities (4.5 km/sec) in the deepest part of the basin (below 2 km), which is in agreement with the P-wave velocities of the deep part of the basin from recent seismic refraction profiling (Snelson et al., 2004) and low velocities (1.5 km/sec) in the shallow basin (200 in). We also performed computations based on the fast marching method approach to solve the forward problem and inversion for basin geometry. This method is used to map the travel-time residual information extracted from the array to variations in subsurface seismic structure. While the coverage of teleseismic data is insufficient to independently resolve the steeply dipping footwall of the basin in its eastern part, we found that the footwall block is likely to be shifted farther west than indicated by the gravimetry-based model. The basin edge is probably related to the Frenchman Mountain fault and its inferred location closer to Las Vegas will result in stronger ground motion during an earthquake. We report site response from teleseismic earthquakes and compare it with previously published site response from regional earthquakes using the standard spectral ratio method. The useful bandwidth of large teleseismic and regional events for standard spectral ratio measurements is 0.1-1.0 and 0.2-5.0 Hz, respectively. Remarkably, we find excellent agreement between the two measurement types within the overlapping frequency band (0.2-1.0 Hz). This indicates that the amplification arises from the structure in the immediate vicinity of the recording station, regardless of the nature of the incoming energy-vertically propagating teleseismic S body waves or horizontally propagating regional surface waves. The results of these investigations indicate that low velocities are present near the surface in LVB, likely related to relatively recent (Quaternary) alluvial and lakebed sediments at the surface. Fast velocities in the deeper basin probably result from older formations.