ALBERT

Description: The Sacramento–San Joaquin Delta is an inland delta at the western extent of the Central Valley. Levees were built around swampy islands starting after the Civil War to reclaim these lands for farming. Various studies show that these levees could fail in concert from shaking from a major local or regional earthquake resulting in salty water from the San Francisco Bay contaminating the water in the Delta. We installed seismographs around the Delta and on levees to assess the contribution of site response to the seismic hazard of the levees. Cone penetrometer testing shows that the upper 10 s of meters of soil in the Delta have shear-wave velocities of about 200 m/s, which would give a strong site response. Seismographs were sited following two strategies: pairs of stations to compare the response of the levees to nearby sites, and a more regional deployment in the Delta. Site response was determined in two different ways: a traditional spectral ratio (TSR) approach of S waves using station BDM of the Berkeley Digital Seismic Net as a reference site, and using SH / SV ratios of noise (or Nakamura’s method). Both estimates usually agree in spectral character for stations whose response is dominated by a resonant peak, but the most obvious peaks in the SH / SV ratios usually are about two-thirds as large as the main peaks in the TSRs. Levee sites typically have large narrow resonances in the site response function compared to sites in the farmland of the Delta. These resonances, at a frequency of about 1–3 Hz, have amplitudes of about 15 with TSR and 10–12 with Nakamura’s method. Sites on farmland in the Delta also have amplifications, but these are typically broader and not as resonant in appearance. Late (slow) Rayleigh waves were recorded at stations in the Delta, have a dominant period of about one second, and are highly monochromatic. Results from a three-station array at the Holland Marina suggest that they have a phase velocity of about 600 m/s and arrive at about the same azimuth as the straight-line back azimuth to the source. A dispersion curve determined for the basin or valley waves yields a shallow velocity profile that increases from about 350 m/s in the upper 0.2 km to about 1.1 km/s at a depth of about 2 km.

Description: We obtain 3D Q P and Q S models for the Delta region of the Sacramento and San Joaquin rivers, a large fluvial–agricultural portion of the Great Valley located between the Sierra Nevada batholith and the San Francisco Bay–Coast Ranges region of active faulting. Path attenuation t * values have been obtained for P and S data from 124 distributed earthquakes, with a longer variable window for S based on the energy integral. We use frequency dependence with an exponent of 0.5, consistent with other studies and weakly favored by the t * S data. A regional initial model was obtained by solving for Q as a function of seismic velocity. In the final model, the Great Valley basin has low Q , with very low Q (〈50) for the shallowest portion of the Delta. There is an underlying strong Q contrast to the ophiolite basement, which is thickest with highest Q under the Sacramento basin, and a change in structure is apparent across the Suisun Bay as a transition to thinner ophiolite. Moderately low Q is found in the upper crust west of the Delta region along the faults in the eastern North Bay area, whereas moderately high Q is found south of the Delta, implying potentially stronger ground motion for earthquake sources to the south. Very low Q values in the shallow crust along parts of the major fault zones may relate to sediment and abundant microfractures. In the lower crust below the San Andreas and Calaveras–Hayward–Rodgers Creek fault zones, the observed low Q is consistent with grain-size reduction in ductile shear zones and is lowest under the San Andreas, which has large cumulative strain. Similarly, moderately low Q in the ductile lower crust of the Bay area block between the major fault zones implies a broad distributed shear zone. Online Material: Distribution of earthquake magnitude, figures of t * values, 3D initial regional velocity and Q models, station terms, and Q P inversion with 1D initial Q P model.