The purpose of this study was to establish three-dimensional GIS-based procedures for evaluating liquefaction susceptibility and to compare the results of these methods to an existing two-dimensional liquefaction susceptibility study for the same area. A recent liquefaction susceptibility study of the Albuquerque area focused on the surficial geology and did not take into account vertical variations of soils that could cause potential liquefaction problems. I developed a GIS-based approach to incorporate these vertical variations by evaluating liquefaction susceptibility for three depth intervals and then combining these to form the overall liquefaction susceptibility for the study area. I also developed a second GIS-based method that used standardized standard penetration test (SPT) blow count data as a proxy for the calculated peak ground acceleration (PGA) trigger to minimize data calculation and processing time. The data for this study are clustered, sparse, and show a high level of spatial variability. These characteristics create some uncertainty with the GIS based liquefaction susceptibility analysis on a regional scale, but would likely not pose a problem with more localized analyses with more evenly distributed data.

The results of these two methods are compared to the previous surficial geology based results. All three methods show the same regional trend, which is highest risk of liquefaction along the river and floodplain with decreasing liquefaction susceptibility to the east and west as the elevation increases and the depth to groundwater increases. Each of the methods is best suited to particular uses, but none of them are without uncertainty or inherent error. The surficial geology approach used in the previous study provides a regional overview of the liquefaction susceptibility distribution, but masks some localized areas were the potential for liquefaction could be higher than the surficial geology indicates, and should primarily be applied to areas with little or no vertical variability. The methods developed for this study provide more localized detail, but when applied on a regional scale introduce uncertainty and bias based on the distribution of the data.