Description |
-
Abstract: Recent earthquakes, such as the 2011 Christchurch, New Zealand earthquake have demonstrated the significant damage potential liquefaction poses to communities. On the regional scale, liquefaction hazard mapping is primarily dependent on the surficial geology to identify liquefaction susceptible units and geotechnical soil data are used to account for the variability within units. However... read more, these studies are costly, therefore not widely available and do not report meaningful probabilities. Thus, there is a need to develop a method of liquefaction hazard mapping that is dependent on globally available geospatial data that can be used in rapid hazard mapping. Logistic regression and ROC curve analysis were used to develop and validate a probabilistic geospatial liquefaction model based on data from the 1995 Hygo-ken Nanbu earthquake in Kobe, Japan. The predicted probabilities of the model successfully captured the spatial distribution of the observed liquefaction features in Kobe. To assess the general applicability of the model, the model was tested in Christchurch, New Zealand and the predictive performance compared to a similar model developed in Christchurch and tested in Kobe. Overall results suggest that on a regional scale, liquefaction occurrences can be assessed using only globally available geospatial data.
Thesis (M.S.)--Tufts University, 2012.
Submitted to the Dept. of Civil Engineering.
Advisor: Laurie Baise.
Committee: Eric Thompson, and Richard Vogel.
Keywords: Civil engineering, and Geological engineering.read less
|
This object is in collection