The Bay Bridges down-hole network consists of 20 sensors that have been placed in bore holes along the Dumbarton, San Mateo, Bay, San Rafael, and Carquinez bridges. Bore holes are between 100 and 1000 ft deep and are drilled 100' into bedrock. Drilling, and lithologic, resistivity and P- and S-wave logging have been provided by Caltrans. The instruments will provide multiple use data that is important to geotechnical, structural engineering, and seismological studies.
Extensive financial support is being contributed by Caltrans, UCB, LBNL, LLNL-LDRD, U.C. Campus/Laboratory Collaboration (CLC) program, and the USGS. The down hole instrument package contains a three component HS-1 seismometer and three orthogonal Wilcoxon 731 accelerometers, and is capable of recording a micro g from local magnitude 1.0 earthquakes to 0.5 g strong ground motion from large Bay Area earthquakes.
The Bay Bridges down hole network consists of seismic sensors in bore holes that are drilled 100 ft into bedrock along and in the San Francisco Bay. Between 1 and 8 instruments have been spaced along the Dumbarton, San Mateo, Bay, San Rafael, and Carquinez bridges. As an example, Figure 12.1 shows a profile of the instrument locations along the Bay Bridge. In addition, two vertical arrays exist at the Dumbarton bridge with additional sensors at the surface and at 200 ft. Two sensors are currently located at the surface at the Bay Bridge and are waiting drill holes. Prior to this study few seismic recording instruments existed in bedrock in San Francisco Bay. This left a recording gap for engineering studies of the Bay bridges and in seismicity studies of the Bay Area. There are six primary areas of research that will be enhanced by the bore hole instrumentation: 1) developing realistic predictions of strong ground motion at multiple input points along long span bridges, 2) examining ground motion variability in bedrock, 3) calibrating soil response models, 4) developing bridge response calculations with multiple support input motions, 5) evaluate the seismicity of potentially active faults in the San Francisco Bay, and 6) record strong ground motion.
Key to these studies is the effort to exploit the information available in weak ground motions (generally from earthquakes with a magnitude of less that 3) to enhance predictions of seismic hazards. Although strong ground motion recordings are essential to calibrate models and understand the hazard of future earthquakes, we can obtain weak ground motion data immediately, whereas it may be years before strong motion data is recorded. The primary research goals utilizing recordings from the Bridges Network are:
Prediction of strong ground motion: LLNL is developing a methodology of using weak ground motion to synthesize linear response strong ground motion and incorporating this with constraints on fault rupture scenarios to predict strong ground motion. These computations provide estimates of the full wavetrain ground motion at multiple points along long span structures.
Ground motion variability: Recent studies have demonstrated the high variability of strong ground motion with site conditions. Recordings along Bay bridges will be used both to improve calculations of ground motions for bridges, and to research the spatial sensitivity and significance of site variability to structures.
Soils response: LLNL is researching means of using weak ground motion to constrain soils models for non-linear computations. Current research has shown that low strain constitutive properties are significant to non-linear ground motion computations, and that these values can be significantly improved by an iterative process of matching weak motion solutions.
Bridge response calculations: Current developments in structural dynamics allow non-linear, three-dimensional calculation of bridge response. This requires realistic full wavetrain input ground motions. LLNL is conducting research on the sensitivity of synthetic ground motions to accurate non-linear computations, and the significance of utilizing multiple support input calculations.
Seismicity: Location of small earthquakes within the Bay that may indicate the existence of active faults will be made possible with the instrumentation. Very small earthquakes (M<2) cannot be recorded adequately to determine accurate locations by regional networks.
Strong ground motion: Strong ground motion from previous earthquakes gives a good indication of what might be expected from future earthquakes. In addition recent earthquakes have demonstrated the high variability of strong ground motion so that an array of strong ground motion recordings will give a better understanding of the ground motion variability from future earthquakes.
The down-hole sensor package is manufactured at LBNL under the direction of Dr. McEvilly, and is the same package used by the USGS and LBNL for the Hayward Fault Digital Recording Network. This package contains three orthogonal Oyo HS-1 4.5 Hz geophones and a three orthogonal Wilcoxon 731s, 10 volts/g, accelerometers. The dynamic range of the Wilcoxon package is from a micro-g to .5 g acceleration, and is flat in frequency response from .1 to 300 Hz. This allows recording of magnitude 1 to 0.5 g strong ground motion from large Bay Area earthquakes. Typically, the Wilcoxon's are recorded over two dynamic ranges to capture weak and strong ground motions, and HS-1's are used as a backup for weak ground motion recording. A Portable Data Acquisition System (DAS) is used to record the data. Sixteen bit resolution Refraction Technology 72A series DAS's are used at most sites with 200 Hz sampling. Three sites utilize Quanterra 4120 24-bit resolution dataloggers with 500 Hz sampling. The data is processed and archived at the BSL.
The deep holes being drilled have been logged by Caltrans to obtain cores of the lithology every 10 feet, an S-wave velocity profile, and a resistivity profile. As an example, Figure 12.2 shows the bore hole log profile for the deep hole at pile 27 on the Dumbarton Bridge. Note that there considerable variability even in the bedrock at depths below 650 feet.
We have had some preliminary results that indicate that the recording of weak motion is extremely valuable. Figure 12.3 shows recordings of acceleration at each of the deep bore hole sensors along the Dumbarton bridge (north component only) from a magnitude 2.1 earthquake located 19 km to the west, on the San Andreas fault. Since the lithology at the bore holes is similar, we might expect the waveforms to be similar at each bore hole from the same earthquake. However, waveforms change significantly at the three deep bore hole sites. There are significant amplitude differences between the recordings, which are not consistent with propagation attenuation effects; and secondary arrivals are present at Pier 27 (center span and center trace in Figure 12.3) that are not present at the two end sites, which is typical of recordings interior to a basin. These same differences will occur in strong ground motion from future earthquakes, and result in phasing and amplitude variations across the bridge.
Uhrhammer, R. A., W. Karavas, and B. Romanowicz, Broadband Seismic Station Installation Guidelines, Seismological Research Letters, 69, 15-26, 1998.