Using H/V measurements to determine depth to bedrock and Vs30 in Boston, Massachusetts

Abstract

Seismic microzonation is an important step in seismic risk and loss studies for densely populated urban regions. Ambient noise studies using H/V measurements to determine the fundamental site period (FSP) can be an inexpensive and efficient method to map site effects for a microzonation study. Ambient noise studies are particularly effective in high impedance contrast environments like Boston, Massachusetts, where the soil depth ranges from 1 to 80m and the soil (shear-wave velocity, Vs=200m/s) overlays hard glacial till or bedrock (Vs=2000–3000m/s). In this study, ambient noise data were collected from 570 locations in the greater Boston area. Nakamura's technique, taking the ratio of the Fast Fourier Transform (FFT) of horizontal components to vertical components, is applied to determine the FSP. The FSP data are mapped across the region and show a consistent pattern with the local geologic and geomorphologic conditions. The FSP data are then paired with depth to bedrock and Vs30 drawn from 2403 boring logs and 25 Vs profiles (Vs30 is defined as the time-weighted average of the Vs in the upper 30m of the soil profile). The depth to bedrock and FSP are correlated, as are FSP and Vs30 measurements. FSP/depth to bedrock and FSP/Vs30 correlations are reported with 95% confidence bounds for use in regional planning. The resulting FSP/Vs30 correlation is comparable with similar studies from high impedance contrast regions in Canada and Japan and show lower Vs30 for sites with FSP>0.2s (where the FSP predicts depth to bedrock>18m) than previously reported for the NGA-West2 database. The accuracy of single station ambient noise results is also confirmed with single-station and vertical-array earthquake spectral ratios. An FSP (~0.74s) was calculated from ambient noise data at the Northeastern University (NEU) Vertical Array site in the Back Bay of Boston. At this site the sediment thickness is 51m and is consistent with the FSP (~0.73s) calculated for the site using earthquake records from the 2011 Mw 5.8 Mineral, Virginia earthquake. Overall, the outcomes of this study demonstrate that ambient noise studies can be used in high impedance contrast environments to reliably and consistently estimate the FSP. When the FSP estimates are paired with additional data such as depth to bedrock, Vs30 estimates, and recorded ground-motion data, the spatial distribution of regional site effects can be characterized.