Surface-wave propagation and phase-velocity structure from observations on the USArray Transportable Array

We address questions relating to the velocity structure of the Earth in three ways: mapping the phase-velocity structure of the western United States, examining deviations of wave paths due to lateral variations in velocity, and demonstrating that Love wave fundamental-mode phase measurements from array methods can be significantly contaminated by overtone interference, dependent on differences in fundamental-mode and first-overtone phase-velocity structure. All of the studies presented in this work use USArray Transportable Array data, which allow for dense, high-quality measurements at an unprecedented level.;To image the uppermost mantle beneath the western US, we improve upon single-station phase measurements by differencing them to produce a baseline data set of phase measurements along inter-station paths, for both Love and Rayleigh waves from 25--100 s. Additional measurements of the arrival angle and local phase velocity are made using a mini-array method similar to beamforming. The arrival-angle measurements are used to correct the two-station baseline measurements and produce a corrected data set. Both the baseline and corrected data sets are separately inverted, producing phase-velocity maps on a 0.5°-by-0.5° grid. We select the corrected maps as the preferred models for Rayleigh waves, with better fits to the data and more consistent measurements. We find that arrival-angle measurements for Love waves may be biased by overtone interference, and hence select the baseline maps as the preferred models for Love waves. The final set of phase-velocity maps is consistent with expectations from known geologic features, and is useful for both calculation of phase for regional paths and studies of radial anisotropy within the region.;We use the mini-array method to make observations of the deviations of waves from the great-circle path. Measured arrival angles vary from 0° to +/-15°. We compile results from earthquakes in small source regions, allowing the observation of bands of arrival-angle anomalies crossing the footprint of the USArray Transportable array in the propagation direction. These bands of deviations may result from heterogeneous velocity structure within the array, or on the larger source-to-array path. We use two global tomographic models to predict arrival-angle anomaly patterns, with both ray-theory-based prediction methods and measurements on synthetic waveforms calculated using SPECFEM3D Globe, a finite element package. We show that both models predict well the long-wavelength patterns of anomalies observed, but not the short-wavelength variations. Experiments with crustal structure indicate that greater heterogeneity is needed in the models. Predictions from the spectral-element-method synthetic waveforms contain the type of complexity seen in the observed patterns, and not obtained with the ray-theoretical methods, indicating that full synthetics are needed to compare model predictions to observed arrival-angle anomalies.;We further examine possible overtone interference in the mini-array arrival-angle and local phase-velocity measurements for Love waves at long periods. Love wave fundamental-mode and higher-mode waves at the same period travel with similar group velocity, making them difficult to separate; the waves have different phase velocities, resulting in a beating interference pattern that oscillates with distance. We show this interference pattern for single-station, two-station, and mini-array phase-velocity measurements. Using measurements on synthetic waveforms calculated using both mode summation and SPECFEM3D Globe, we show that contamination of single-station measurements can largely be explained by interference between the fundamental and first-higher mode only. Interference causes small variations in the single-station phase velocity, up to 1%, and the oscillations about the expected values are asymmetric. The two array-based measurement techniques can be thought of as a spatial gradient over the single-station phase measurements, and consequently much larger variations are observed, 10--20%, and the results are biased to higher phase velocities. We conclude that overtone contamination must be carefully considered prior to attributing array-based Love wave phase-velocity anomalies to Earth structure.