Evaluation of Subsurface Impacts on Occurrence of Sand Boils from the 2001 Nisqually Earthquake in the SoDo Neighborhood of Seattle, Washington

Abstract

The 6.8 magnitude Nisqually earthquake occurred on February 28th, 2001 causing liquefaction in multiple areas of Seattle, mostly in the SoDo neighborhood. Following the earthquake, scientists recorded field observations of the liquefaction in Seattle, which was compiled into a database. While the SoDo neighborhood is mapped as having the same liquefaction potential throughout, large areas did not show above-ground signs of liquefaction, such as sand boils. I looked at the subsurface in areas where sand boils were reported and in areas where no sand boils were reported to determine if significant geologic differences occur in the subsurface that could explain why liquefaction occurred in some areas and not others. The study area has undergone multiple glaciations, and is part of the former Duwamish River delta, occupying a former subglacial trough. In the last century, humans have modified the delta and associated tide flats to expand buildable land by adding variable thicknesses of fill, primarily sourced from nearby glacial deposits. To examine the subsurface, I used the GeoMapNW subsurface database to find borings adjacent to sand boils and borings at sites where sand boils were not reported. I identified target locations, where sand boils did and did not occur, and where good subsurface data are available. I compared the subsurface stratigraphy to depths of 40 feet from selected borings in the target areas and looked at density, sand content and clay content. I also compared the ratio of the thickness of fine-grained sediment to coarse-grained sediment. To compare the densities of the subsurface, I performed clean sand corrections and performed modeling to generate a Liquefaction Potential Index (LPI) at the target locations. I found that areas with capping clay layers at/near the surface of greater than 3ft in thickness or significant amounts of clay within the upper 40 feet were more common in areas where no sand boils were reported. I also found that areas where no sand boils were reported typically had higher ratios of fine-grained thickness to total thickness, and that these ratios were higher in the eastern portion of the site area. The LPIs at borings adjacent to sand boils were calculated to be high risk or very high risk. The LPIs for the borings chosen in areas where there were no sand boils ranged from low risk to very high risk. Some of the high or very high risk boring locations had caps of clay or near surface layers had undergone construction-related compaction, which may have resulted in liquefaction not reaching the surface and instead potentially spreading laterally. For the Nisqually earthquake the LPI alone does not adequately predict whether or not liquefaction will occur at the ground surface in the SoDo area due to geologic or human-influenced variability. LPI will over predict liquefaction at the ground surface for the SoDo area.