Abstract: Observational seismology plays a crucial role in advancing our understanding of the Earth's dynamic processes and internal structure, serving as the foundation of seismic hazard assessment. It relies heavily on the availability and quality of data from a wide range of sources. Distributed Acoustic Sensing (DAS) is an emerging technology with the potential to greatly expand seismic data coverage by converting fiber-optic cables into dense arrays of seismic sensors. Compared to conventional instruments, DAS offers unique advantages in spatial density and convenient deployment, particularly in challenging or previously inaccessible environments. However, DAS also presents several limitations, including lower signal-to-noise ratios for individual channels, indirect measurements of ground motion, and directional sensitivity to axial fiber orientation. As a result, data processing procedures for routine seismic monitoring need to accommodate these features. This thesis contributes to developing modified processing techniques and evaluating the performance of DAS arrays under various conditions across three key applications: event detection, source imaging, and shallow subsurface characterization. The findings of these case studies aim to provide implications to the seismological community in assessing the potential for integrating DAS into modern seismic networks.

In Chapter 2, we focused on assessing the recording capability of an Ocean-Bottom DAS (OBDAS) array in the Sanriku region, Japan. We introduced two array-based detection methods—Waveform Similarity Search (WSS) and Spectrum Similarity Search (SSS)—that utilize the dense spatial sampling of OBDAS to detect coherent earthquake signals over subsections of the array. These techniques detected thousands of cataloged and previously uncataloged earthquakes. By analyzing the detection statistics, We found that the recording capability of the OBDAS array varies substantially across channels, and the array is well capable of recording regional earthquakes within a 100 km radius region. The array also recorded local repeating earthquakes across different subregions. These results highlight the feasibility of using OBDAS for long-term seismic monitoring and its potential to address the scarcity of offshore instrumentation, especially near subduction zones with extensive submarine fiber networks.

In Chapter 3, we investigated the potential of DAS on earthquake rupture imaging. We utilized both synthetic data and realistic recordings to identify the significant challenges of applying the Back-projection method (BP) to DAS data: the unstable solvability caused by highly asymmetric array geometry and limited azimuth coverage. Considering these constraints, we also proposed several data processing procedures to better adapt DAS data for BP analysis. We demonstrated the effectiveness of BP with the 2022 MW7.6 Michoacán earthquake recorded by a DAS array in Mexico City. Our analysis demonstrated that, despite some limitations, DAS-based BP could successfully capture key rupture features, including multiple subevents and rupture direction. Meanwhile, we analyzed several sources of uncertainty and proposed practical guidelines for improving DAS-based BP performance. We also proposed an initial assessment scheme to understand the feasibility of BP analysis for a given event-array geometry, which is transferrable to other similar studies. Our work highlights the potential of DAS to enhance earthquake source imaging on a regional-to-local scale, offering alternative yet valuable insights into regions under-served by conventional seismic networks.

In Chapter 4, we used ambient seismic fields recorded by an ocean-bottom DAS array to image the shallow subsurface beneath the Florence region. Leveraging the long-duration recordings of DAS, we retrieved coherent surface waves through extensive stacking and applied a double-beamforming (DBF) approach to stably measure multimode dispersions. We performed a perturbational-based inversion method to invert for S-wave velocities over the first 2000-meter sediments underlying the fiber-optic cable. Our results demonstrate the effectiveness of the DBF method in enhancing spatial resolution. While the high cost and limited availability of underwater instruments hinder progress in imaging shallow structures in marine settings, this work demonstrates the potential of OBDAS arrays for high-resolution passive imaging.