Keywords

Horizontal-to-Vertical Spectral Ratio, Shear-Wave Velocity, Peak Ground Acceleration, Fundamental Frequency, Peak Amplitude, Surface-to-Borehole Spectral Ratio

Abstract

The prevailing seismic site classification parameter, the time-averaged shear-wave velocity in the upper 30 meters (VS30), provides an incomplete representation of site conditions, neglecting deeper geology, impedance contrasts, and resonance phenomena. This thesis developed and validated a data-driven framework for seismic site classification based on the full spectral shape of Horizontal-to-Vertical Spectral Ratio (HVSR) curves combined with unsupervised machine learning, through two complementary research studies. Research 1 applied K-means clustering to earthquake-based HVSR (eHVSR) data from 3,186 ground motion stations across California and adjacent regions, identifying four physically distinct site clusters, characterized by low-frequency, very-low-frequency, high-frequency, and flat/multi-peak spectral signatures. Comparison with conventional VS30-based site classification results revealed that clusters with nearly identical median VS30 values could have different spectral shapes, confirming that VS30 cannot fully represent frequency-dependent resonance behavior and deep stratigraphic influences. Research 2 extended the clustering method to a continental scale using microtremor HVSR (mHVSR) data from 5,206 quality-controlled stations across the Contiguous United States. A multi-criteria evaluation identified seven spectral clusters with three distinct seismic site classes: Class D (Deep Sediment Sites), Class M (Intermediate Sites), and Class S (Shallow Sediment Sites), with seven sub-classes. Independent validation with shear-wave velocity profiles confirmed statistically distinct subsurface signatures across the classes, characterized by systematic differences in velocity gradient, depth to engineering bedrock, and maximum velocity contrast. Further validation using regional sediment thickness and VS30 datasets supported the geological correspondence of the proposed classes and revealed significant overlap of VS30 across sub-classes, highlighting the advantage of full-spectrum classification. Collectively, these studies advance the non-invasive HVSR technique in seismic hazard assessment by providing a new data-driven and physically plausible seismic site classification method.

Completion Date

2026

Semester

Spring

Committee Chair

Zhan, Weiwei

Degree

Master of Science in Civil Engineering (M.S.C.E.)

College

College of Engineering and Computer Science

Department

Civil, Environmental and Construction Engineering

Document Type

Dissertation/Thesis

Identifier

DP0053212

Download

Share

COinS
 

Accessibility Statement

This item was created or digitized prior to April 24, 2027, or is a reproduction of legacy media created before that date. It is preserved in its original, unmodified state specifically for research, reference, or historical recordkeeping. In accordance with the ADA Title II Final Rule, the University Libraries provides accessible versions of archival materials upon request. To request an accommodation for this item, please submit an accessibility request form.