1997 SCEC Press Report

Project Title: Completion of Phase III and Beyond
PI's: Jamison H. Steidl and Alexei G. Tumarkin
Institution: Institute for Crustal Studies, UC Santa Barbara

Phase III work

The focus of the site response section of the Phase III report has been on looking for measurable and/or mapable parameters that can be used to reduce the uncertainty in predicting ground motion when using attenuation relations. Using both detailed and general digital geologic maps, measured shear-wave velocity data, and a depth to basement parameter, we examine the correlation between these parameters and empirical strong- and weak-motion data recorded in Southern California. The results from this work can then be used to incorporate site response into hazard maps.

As expected, a correlation between geology and site response is found in the response spectral acceleration of strong motion data for Southern California. Older more competent sediment or bedrock show less amplification of ground motion than younger less consolidated alluvium. In order to incorporate site response into the hazard calculations, we calculate response spectral amplification factors relative to a rock attenuation relation for specific geologic site classes. Previous results suggest that more detailed maps of Quaternary age sediments may provide useful information for site response studies. However, in order to use results on a regional mapping scale we are forced to use the more general QTM state map of California (Q=Quaternary, T=Tertiary, M=Mesozoic) since more detailed digital maps are not available for the whole Southern California region.

The strong-motion data used in this study are response spectral acceleration at periods of 0.1, 0.3, 1.0 and 3.0 seconds, taken from the SCEC strong-motion database (SMDB). We used data for all earthquakes from 1933 to present at all stations with locations between 32o and 36o North latitude. We excluded stations in buildings with two stories or more and dam abutments. - At each period we estimated the site response for a given recording by comparing it with the ground motions expected at a rock site at the same distance from the given earthquake. The comparison is made by computing the ratio of the recorded motion to the reference rock motion, where the Sadigh et al. (1994) rock attenuation relation is used as the reference.

There is a large variability in the individual ratios for a given site class with a trend to increased variability at longer periods. This is most likely due to the different depths of soil profiles at each site, some deep enough to affect the long period motion, and others shallow enough that the long period waves don't even see the shallow soil layers. Weak-motion response spectral levels from dense array data recorded during the 1994 Northridge portable deployment show large variability for the same site class (Field and Hough, 1997), so the variability we see in the strong-motion response spectral acceleration is not surprising. In addition, the event to event variability due to the uniqueness of each earthquake source adds to the variability in the ratios. In order to make these ratios useful for seismic hazard maps we need to look at what the ratios (site response) are (is)
doing on average, and assign a mean value and standard deviation to each site class. The averaging of the site response for each site class is described below and shown in Table 1.

The ratios mentioned above, or individual amplification factors, are averaged for each site class, Q, T, and M, after being separated into bins of similar rock input PGA (peak ground acceleration). The reason for separation into bins of similar rock input PGA is to allow for any nonlinearity of the site response as the ground motion increases. The average amplification factors are given in Table 1 for each site class, each period, and each input rock PGA bin. The main conclusions from these results are an increase in amplification in going from older to younger site class, and a decrease in amplification in going from smaller to larger rock input PGA.

In the example shown in Figure 1, we use the 3.0 second period values from Table 1 and a correction based on the depth to basement to see how we would do in predicting the 3.0 second ground motion from the January 17, 1994 Northridge earthquake. The depth to basement relation, ln(A)=-2.1+0.321n(D), where A is the amplification of 3.0 second period ground motion in addition to any site corrections from Table 1, was determined from all the 3.0 second Southern California strong-motion data. The depth to basement parameter D is taken from the model of Magistrale et. al (1996) and is the depth to the 2.5 km/s shear-wave velocity isosurface. Each point in Figure 1 represents a strong-motion station that recorded the Northridge earthquake. We plot the residual ln(observed) - ln (predicted) versus distance from the fault surface. The prediction uses the rock attenuation relation with the site correction (Table 1) and basin correction above
(Figure 1A), and a 3-D synthetic calculation (Olsen and Archuleta, 1996) of the 3.0 second response (Figure 1B).

The example in Figure 1 shows that in both the 3-D synthetic simulation and the empirical attenuation calculation (with corrections), there is a trend to overpredict as we move away from the fault. The mean value taken over all distances shows an small underprediction for the 3-D synthetic simulation and a larger overprediction for the empirical calculation. The 3-D synthetic model uses a realistic finite fault source model, while the empirical attenuation calculation does not, which could account for the differences in the mean. The standard deviation is slightly smaller for the empirical calculation. In general, both methods are able to predict the ground motion to within a factor of three, with the majority of sites predicted to within a factor of two.

Portable uphole/downhole instrumentatzon

In addition to the work towards completion of the Phase III report, in the 1997 fiscal year we began an experiment to take a closer look at the effect of the shear-wave velocity in the upper 30 meters on site response. This is a very critical issue since the Uniform Building Code (UBC) is using shear-wave velocity in the upper 30 meters as a means to define site class. Previous results shown in last years SCEC progress report show that there is a large variability in the correlation between site amplification and shear-wave velocity in the upper 30 meters. In order to examine more closely the relationship between near-surface shear-wave velocity and ground motion amplification, we have purchased a 2.5" diameter borehole instrument package which uses a Wilcoxon accelerometer. This package is to deploy in multiple 3" diameter cased holes left by the USGS after logging for velocity at the Los Angeles Dam complex in the San Fernando valley, and other sites throughout Los Angeles. An inflatable coupling system used to secure the instrument package in the borehole has been designed and makes the system portable, so that we can retrieve the package and move on to the next borehole. Weak-motion data from local earthquakes will be used to analyze the contribution of the upper 30 meters on site response at multiple sites. The permitting phase of this project has been completed with the Los Angeles Department of Water and Power, and the Metropolitan Water District. Delays in the shipment of connectors for the downhole package power supply has pushed the first installation to December 1997.
References

Field, E. H., and S. E. Hough (1997). The variability of PSV response spectra across a dense array deployed during the Northridge aftershock sequence, Earthquake Spectra, (in press).

Magistrale, H., K. McLaughlin, and S. Day (1996). A geology-based 3D model of the Los Angeles basin sediments, Bull. Seism. Soc. Am., 86, pp 1161-1166.

Olsen, K. B. and R. J. Archuleta (1996). Three-dimensional simulation of

o earthquakes on the Los Angeles fault system, Bull. Seism. Soc. Am., 86, pp 575-596.

Sadigh, K. R., C.-Y. Chang, N. A. Abrahamson, S. J. Chiou, and M. S. Power (1994). Specification of long-period ground motion: Updated attenuation relationships for rock site conditions and adjustment factors for near-fault effects, Proceedings of the International Workshop on Strong Motion Data, Vol. 2, pp 237-248, Menlo Park, CA.

SCEC Publications

Bonilla, L. F., J. H. Steidl, G. T. Lindley, A. G. Tumarkin, and R. J. Archuleta (1997). Site amplification in the San Fernando Valley, California: Variability of site-effect estimation using the S-wave, coda, and H/V methods, Bull. Seism. Soc. Am., 87, No. 3, pp. 710-730.

Bonilla, L. F., J. H. Steidl, and A. G. Tumarkin (1996). Site amplification in the San Fernando Valley from weak- and strong-motion data, Paper No. 1634, Proceedings of the Eleuenth World Conference on Earthquake Engineering, Acapulco, Mexico, Elsevier.

Steidl, J. H., A. G. Tumarkin, and R. J. Archuleta (1996). What is a reference site? Bulletin of the SeismologicaZ Society of America, 86, No. 6, pp. 1733-1748.