S-Wave Velocities - 1999 Report
S-wave Velocities for Specific Near-Surface Deposits in Seattle, Washington
Poster presented at 1999 Spring Meeting of the SSA
Williams, R.A., Stephenson W.J., Odum J.K., and Worley, D.M, 1999, Seismological Research Letters, v. 70, no. 2, p. 257.; rawilliams@usgs.gov
Abstract
We determined the compressional- and shear-wave velocities (Vp and Vs) to about 40-m depth at 25 locations in the Seattle urban area (see right) from high-resolution P- and S-wave seismic-refraction and reflection data. Seventeen sites were located where portable digital seismographs recently recorded earthquakes or airgun shots from the 1998 SHIPS experiment. These data identify distinct Vs for near-surface deposits including the Pleistocene Lawton Clay (300-400 m/s), Esperance Sand (~400 m/s), and Vashon Till (600-900 m/s) and help explain observed variations in earthquake ground motions (site response). From our measurements, we determined that a near-surface Vs inversion exists in areas where Vashon Till overlies Esperance Sand and Lawton Clay. The association of a distinct Vs with specific surficial deposits allows extrapolation of site response to areas where Vs has not been measured. Sites with the lowest measured Vs correlate with highest ground motion amplification in the 1-10 Hz frequency band, relative to a reference rock site. These sites, such as at Harbor Island and in the industrial area south of the Kingdome, are located on artificial fill and have an average Vs in the upper 30 m of 150-170 m/s. Such low S-wave velocities classify these sites as NEHRP soil profile type E (average Vs less than 180 m/s). The Vs of the Blakely Formation sandstone is much higher in the upper 30 m in West Seattle at Alki Point (about 1200 m/s) than at Seward Park (about 500 m/s). This velocity structure would produce a higher impedance contrast between bedrock and the overlying Pleistocene deposits in West Seattle than in the vicinity of Seward Park and may explain part of the increased site response observed in West Seattle during earthquakes (Frankel et al., 1999 and Carver et al., 1999).
(Click on map for larger image)
Project Summary
The near-surface site conditions are critical parameters needed to understand and forecast the amount of earthquake ground motion in and urban area. For in-situ determination of near-surface P- and S-wave seismic velocities, high-resolution seismic reflection/refraction methods, operated from the ground surface, have proven to be a cost-effective alternative to drilling a borehole and conducting a downhole seismic survey. We have acquired near-surface data using surface-based techniques at about 50 sites in the past 7 years (Williams and others, 1994; 1996; 1997; Carver and others, 1998 ). These techniques, which are non-invasive, unobtrusive, and offer relatively high surface coverage (averaging two sites per day), make them ideal for inner-city microzonation studies. Additionally, we are working to further standardize and simplify the data acquisition and processing procedures so that the method is an accepted and useful practice by the earthquake engineering community.
The objective is to provide site response and ground motion modeling support in the Puget Lowland by acquiring high-resolution seismic reflection/refraction data. This is achieved by determining near-surface (0 to 30 m) P- and S-wave seismic velocities at needed localities and earthquake recording sites. These data will show whether differences in site response can be partially explained by different near-surface P- and S-wave velocity structure, and if mapped geologic units can be distinguished by these methods.
Method
The seismic profiles were acquired on the paved streets or city parks in residential areas, or railyards and any open ground in industrial areas (see photo right). We interpreted the data using the slope-intercept method of analysis (Mooney, 1984; see figure). Recording parameters are listed in table 1.
Table 1.
Seismic-Refraction/Reflection Data Recording Parameters
Recording system: Geometrics StrataView (30 chan. P-wave, 60 chan. S-wave)
Sampling interval: 0.001 seconds
Record length: 1 second
Recording format: SEG-2
Geophones: Sixty 4.5-Hz horizontal; thirty 8-Hz vertical
Geophone array: linear, with single phones at 1.5-m intervals (S-wave) and 3.0-m intervals (P-wave)
Source: 4.0 kg sledgehammer on metal plate (P-wave); 4.0-kg sledgehammer on wood timber (S-wave)
Source array: Reversed spreads with multiple off-end shots
Reversed seismic S-wave profiles ranged in length from 87 to 177 m. These S-wave profile lengths resulted in a maximum survey depth range of about 30 m. The S-wave seismic source consisted of a wooden timber placed on the pavement beneath the wheels of the vehicle at right angles to the direction of the profile (see photo below). Reversed polarity seismic energy was produced by striking opposite ends of the timber with a 4-kg sledgehammer. We picked first-arrival phases assumed to be refracted from the same interface, calculated the velocity from the slope of the line connecting these phases, and then extended the line connecting these phases back to the zero offset point. We determined that the slopes were accurate to within about 5-10 percent. Thus the calculated layer thicknesses had roughly the same accuracy. There are two limitations underlying this technique: (1) an assumption that layer velocity is constant across the length of the profile, and (2) low-velocity layers underlying a high-velocity layer cannot be detected. In spite of these assumptions and level of accuracy, this approach has been shown to generally agree with seismic velocity downhole profiles determined from shallow boreholes in the Los Angeles area (Williams et al., 1996, 1997).
Seismic Profiles
not measured
Quaternary, older unconsolidated- not measured for seismic velocity
Results
A comparison of average horizontal amplitude from 2-6 Hz (blue), the peak horizontal amplitude from 0.5-20 Hz (light yellow), the frequency at which the peak amplitude occurs (maroon), and the average S-wave velocity to 30-m depth at these sites (aqua). The S-wave velocity has been normalized by dividing by 100. The chart shows that low site response is associated with higher S-wave velocities with the exception of the two sites in West Seattle (WEN and WES). The peak amplitudes are shown to emphasize that averaging amplitudes can sometimes obscure strong resonances that are up to a factor of 2 greater than the average amplification factors. Near-surface impedance boundaries responsible for strong resonances have been directly detected in Seattle and the central U.S. by Williams et al. (1998) using seismic reflection data. An example of one of theses reflectors that generates a resonance at about 0.9 Hz at site VMF in Seattle is shown in the middle seismic section of the yellow-shaded zone to the upper left.
Average horizontal spectral site amplification from 2-6 Hz from Frankel et al. (1999) and Carver et al. (1999) versus mean shear-wave velocity to 30-m depth. The red line fits all data and the blue line fits the data except for the two West Seattle sites. The site amplification values were determined from weak-motion earthquake data, but weak-motion data like this has been used to successfully predict strong ground motion in the San Francisco Bay Area (Borcherdt and Glassmoyer, 1994). The data show that site amplification in this frequency band is inversely proportional to the square root of the shear-wave velocity.
