We develop several new techniques to better retrieve Earth's structure by analyzing seismic surface waves. These techniques are applied in regional studies to understand a variety of tectonic structures and geodynamic processes in Earth's crust and upper mantle. We create an automated method to retrieve surface-wave phase velocities using dense seismic arrays. The method is based on the notion of using cross-correlation to measure phase variations between nearby stations. Frequency-dependent apparent phase velocities are inverted from the phase-variation measurements via the Eikonal equation. The multi-pathing interference is corrected using amplitude measurements via the Helmholtz equation. The coherence between nearby-station waveforms, together with other data-selection criteria, helps to automate the entire process. We build up the Automated Surface-Wave Measuring System (ASWMS) that retrieves structural phase velocity directly from raw seismic waveforms for individual earthquakes without human intervention.

This system is applied on the broad-band seismic data recorded by the USArray from 2006-2014, and obtain Rayleigh-wave phase-velocity maps at the periods of 20-100~s. In total around half million seismograms from 850 events are processed, generating about 4 million cross-correlation measurements. The maps correlate well with several published studies, including ambient-noise results at high frequency. At all frequencies, a significant contrast in Rayleigh-wave phase velocity between the tectonically active western US and the stable eastern US can be observed, with the phase-velocity variations in the western US being 1-2 times greater. The Love wave phase-velocity maps are also calculated. We find that overtone interference may produce systematic bias for the Love-wave phase-velocity measurements. We apply surface-wave analysis on the data collected by a temporary broad-band seismic array near the D'Entrecasteaux Island (DI), Papua New Guinea. The array comprises 31 inland and 8 off-shore broad-band seismic sensors, and were operated from March 2010 to July 2011.

We adopt the ASWMS to retrieve phase velocities from earthquake signals, and apply the ambient-noise analysis to obtain the Rayleigh-wave phase velocities at higher frequencies. The multi-band phase velocities are inverted for a three-dimensional shear-velocity model of the crust and the upper mantle. The result reveals localized lithosphere extension along a rift-like axis beneath the DI, with a shear-velocity structure similar to an adiabatic upwelling mantle. West of the DI, very slow shear velocities are observed at shallow mantle depth (30-60~km), which we interpret either as the presence of in situ partial melt due to inhibited melt extraction, or as the existence of un-exhumed felsic crustal material embedded within the surrounding mantle. Love waves contain important information to constrain the upper-mantle radial anisotropy. However, Love-wave fundamental-mode phase-velocity measurements are often contaminated by overtone interference, especially within regional-scale arrays.

We evaluate this problem by analytically and numerically evaluating the behavior of synthetic wavefields consisting of two interfering plane waves with distinct phase velocities but comparable group velocities. The results indicate large phase variance due to the interference that can explain the systemic bias observed in data. We develop a procedure that utilizes amplitude measurements to correct for the interference effect. The synthetic tests show the correction can significantly reduce the phase-velocity variance and the bias generated by the interference.