Date of Award


Document Type

Open Access Dissertation


Civil and Environmental Engineering


Civil Engineering

First Advisor

Sarah Gassman


Seismically-induced soil liquefaction is one the most hazardous geotechnical phenomenon that can cause loss of life and devastating damage to infrastructure. Proper estimation of critical ground motion parameters (e.g. peak ground acceleration and earthquake magnitude) is vital for seismic design of new structures and retrofit of existing structures, especially in regions such as the South Carolina Coastal Plain (SCCP) where the frequency of re-occurrence of large earthquakes is low (studies of paleoliquefaction features have revealed seven, large, prehistoric earthquakes occurring within the last 6000 years) and the locations of potential sources are not exactly known. Moreover, due to mechanical and chemical mechanisms, phenomena known as “aging”, soil resistance to liquefaction increases with time and so the age of soil deposition must be considered in liquefaction analysis of aged soil deposits in the SCCP.

In 2005, a method was developed to estimate the minimum earthquake magnitude, M, and peak ground acceleration, amax, of prehistoric earthquakes using in-situ geotechnical data (e.g. cone penetration and standard penetration data), back-calculation methods (e.g. the Energy Stress and Cyclic Stress methods) and several approaches that account for soil aging by considering density changes in the soil with time. Since then, newer semi-empirical approaches have been published based on an expanded case history database of liquefaction/no liquefaction sites. Ground Motion Prediction Equations (GMPEs) have also been used in combination with back-calculation methods to estimate the earthquake magnitudes. Additional studies related to soil age have also been published. Therefore, the purpose of this study is to use these newer approaches to further improve the current estimates of M and amax at the four sites of Hollywood, Fort Dorchester, Sampit, and Gapway located in the SCCP.

The first study in this dissertation improved upon the 2005 study by using a newer semi-empirical liquefaction analysis method to update the cyclic resistance ratio (CRR) and back-calculate the minimum peak ground acceleration at Sampit and Gapway sites. The effect of aging on soil resistance was taken into account using the same methodology as in 2005. Results show that the newer method for calculating CRR produces lower peak ground accelerations than the previously used approach. The difference is most significant for lower magnitudes. Calculated average values of age-adjusted magnitude range from 5 to 7.5 and the corresponding age-adjusted peak ground acceleration range from 0.08 to 0.23g.

The newer method used in the first study was also used at the Hollywood site, a site that had not been previously studied in 2005, and has evidence of four episodes of paleoliquefaction. The results are presented in the second study of this dissertation. For the Hollywood site, it was shown that when the age of the earthquake was not considered, the magnitude ranged from 7 to 7.2 and the corresponding acceleration ranged from 0.23 to 0.35g. The minimum earthquake magnitude at the time of earthquake was found to be lower when accounting for age. As an example, for the most recent prehistoric earthquake with the age of 546±17, the minimum back-calculated magnitude ranged from 5.7 to 6.7 with corresponding acceleration ranging from 0.17 to 0.30g.

The third study of this dissertation used a newer aging approach that considers the influence of age, cementation and stress history on the CRR of the soil to back-calculate the minimum earthquake magnitudes at the Fort Dorchester site. The new aging approach provided magnitudes that ranged from 5.1 to 6.2 and were in general agreement with previously used methods that considered the effect of aging on only the CPT tip resistance values. Also, when the size of the fault was considered, the maximum magnitude was found to be 5.6 and the corresponding peak ground acceleration ranged from 0.21 to 0.36 g.

The fourth study of this dissertation presents the results from a statistical analysis performed on the available geotechnical data set to find a relation between the updated obtained cyclic resistance ratio values. Significant correlation between equivalent clean sand tip resistance and the cyclic resistance ratio at the time of earthquake was shown using descriptive statistics, summary statistics and regression analysis on the current measurements of field test data.

The fifth study of this dissertation used four proper GMPEs for the east coast of the US combined with the Cyclic Stress method to predict the minimum earthquake magnitude and peak ground acceleration at the Hollywood, Fort Dorchester, Sampit and Gapway sites and find a regional assessment of amax-M in the SCCP. Results were compared with previously found values using the Cyclic Stress and Energy Stress methods. It was shown that when the source of the earthquake is associated with the Charleston Source, the minimum earthquake magnitudes for the prehistoric earthquakes that occurred between about 546 to 1021 years ago and between 3548 to 5038 years ago were estimated to range from 6.6 to 7.5 and 6.1 to 7.2, respectively. For the earthquakes associated with the Sawmill Branch Fault that occurred about 3500 years ago or earlier, the minimum earthquake magnitudes were estimated to range from 5 to 6.3.