AN ANALYSIS OF SIMULATED AND OBSERVED BLAST RECORDS IN THE SALT LAKE BASIN

We have simulated 0.2 to 1.2-Hz 3D elastic wave propagation in the Sal t Lake Basin from a blast at a nearby open-pit mine. A fourth-order st aggered-grid finite-difference method was used to simulate the blast i n a two-layer basin model (58 x 43 x 9 km) consisting of semiconsolida ted sediments up to 1.3-km thick surrounded by bedrock. Data from four blasts in the mine pit, recorded by a network of 10 digital three-com ponent instruments, were compared to the results of the simulation. Th e simulation reproduces the overall pattern of ground-motion amplifica tion at basin sites relative to a rock site, as measured by ratios of peak particle velocities, cumulative kinetic energies, and spectral ma gnitudes. Considering the simple two-layer basin model used in the 3D simulation, this finding suggests that the deep 3D basin structure sig nificantly contributes to low-frequency ground-motion amplification in the Salt Lake Basin. Order-of-magnitude discrepancies exist between s ome of the observed and predicted ground-motion parameters, and the si mulations underpredict the signal durations at most stations. We use 2 D simulations along a profile through the southern part of the basin m odel to investigate the causes of these discrepancies. These causes ma y be summarized, in order of their importance along this profile, as f ollows: 1. Effects of a near-surface layer of low-velocity unconsolida ted sediments (P- and S-wave velocities of 1.65 and 0.41 km/sec, respe ctively) that at soil sites along the profile increase the peak partic le velocities by up to a factor of 3 and significantly increase the gr ound-motion durations. 2. Attenuation in the sediments, which greatly diminishes the ground-motion durations on the synthetic seismograms wh en parameterized by realistic values of the quality factor, Q (20 for soil sites and 35 for bedrock sites). 3. 2D topographic scattering, wh ich increases the peak particle velocities by up to a factor of 2 and increases the signal durations for sites along the profile. Compared t o the records from the simple two-layer 3D simulation, the records fro m a 2D P/SV-wave simulation that includes processes (1) through (3) pr ovide a better match to the blast data-especially the observed duratio ns of shaking. At five of the six stations along the profile, the 2D s imulation reproduces the normalized radial and vertical peak particle velocities to within a factor of 2 and the normalized cumulative kinet ic energies and spectral amplitudes on these components to within gene rally a factor of 3. Our results suggest that deep-basin resonance, re verberations in the near-surface low-velocity layer, attenuation, and topographic scattering significantly influence site amplification in t he Salt Lake Basin. Future studies of site amplification in the Salt L ake Basin should include the effects of all of these mechanisms.