Annual Report 1997

Analysis of Northridge Aftershock Amplitudes and Damage

Paul Davis, Liu Hong, Shangxing Gao, and Monica Kohler

Group B.

Our group B work in the last year can be separated into 3 categories:
1. Continued study of Northridge aftershock data.
2. Site selection and planning/proposal writing for a vibroseis/explosion survey of Santa Monica
damage zone.
3. Installation of 18 stations across LA basin to search for amplification/focusing effects in local,
regional and teleseismic data.
A brief description of highlights from each category follows:

la. Origin of Enhanced Damage in Sherman Oaks

During this year Hong Liu concentrated on analyzing the Northridge aftershock data from the Sherman Oaks region for comparison with the Santa Monica region (Gao et al., 1996).

In the south San Fernando Valley (SFV) at the time of the Northridge earthquake, Sherman Oaks area was perhaps the most heavily damaged area in terms of the percentage of redtagged buildings. The largest density of red-tagged buildings is concentrated in an E-W zone between a narrow E-W strip of old sediments to the north and the southern boundary of the SFV to the south (Figure 1). The overall density of buildings in the southern part of Sherman Oaks, defined here as the area between Mulholland Drive and the valley boundary, is much lower than that in the northern part, which includes the area between the valley boundary and the strip of old sediments. Therefore, the number of red-tagged buildings may not be a proper indicator of the amplitude of the mainshock in Sherman Oaks. Given the difference in the overall density of buildings, it is conjectured that the mainshock amplitudes were roughly comparable in the northern and southern parts. Possible causes for the enhanced damage include the following surface geology, liquefaction and basin focusing. Given the variable geology we conclude that basin focusing played a sign)ficant role.

Basin focusing One of the diagnostics of basin focusing is epicentral dependence of peak amplitudes of P and S waves. We use the following procedure to quantify such a dependence. For each event, the mean amplitude for the records from stations in the SM mountains and another mean amplitude for Sherman Oaks are calculated and a ratio between the two is obtained. The ratio is plotted on a polar plot as a function of incidence and azimuth of the incoming rays. Figure 2 shows results for the S wave peak amplitudes, and Figure 3 for the P waves. A strong azimuthal dependence of amplification factors for both P and S waves is obtained. The amplitudes for events from the north are about 3 times larger than those from the NW.
The fact that rays from the north are more strongly amplified may imply that the strike of the structure that focuses them is E-W. One of the candidates for such a structure is a lens formed by the southern flank of the older sediment strip and the southern boundary of the SFV. However, the fact that stations to the south of the boundary have higher amplitudes suggests that the southern boundary of the lens-structure is a deeper discontinuity separating the basement and the sediments (Figure 1).

We have constructed a two-dimensional model for Sherman Oaks based on the local geology. Figure (4) shows the southern boundary of the San Fernando Valley, and the crosssection of a high-velocity sedimentary 'island'. Enhanced damage was observed in the small basin formed by the southern boundary of the island and the Valley boundary. The vertical extent of the small island is not well-defined, and therefore it is difficult to make a direct comparison between the calculated and observed seismograms. Snapshots for SH waves are shown in Figure 5, where a strong focusing effect is clearly observed.

lb. Estimation of Strong Motion in the Central Damage Zone of Santa Monica

In last year's report we presented regional response spectra of P and S waves in aftershock data from our Northridge Earthquake array. Figure 2 from that report shows that the focusing phenomenon, inferred to have caused enhanced damage in Santa Monica, is frequency dependent, and occurs mainly for frequencies greater than 3 Hz. We have attempted to use this transfer function to estimate how the Santa Monica City Hall (SMC) strong motion record would have appeared at the center of the damage zone.

The (SMC) strong motion seismometer was located within the damage zone, but focusing effects, though present, were not at their maximum. The acceleration record is dominated by a strong high frequency (5 Hz) burst of acceleration peaking at about 0.9g followed by smaller lower frequency (0. 5 Hz) waves.

In order to estimate strong ground motion at the center of focusing we calculated the transfer function that converted the spectra of nearby aftershock stations to that of a station at the center of the focusing zone. We then multiplied the spectrum of the strong motion data by this transfer function and transformed back into the time domain. Because the lens effect selectively amplifies high frequencies, the high frequency accelerations at SMC which had a peak of 0.9 g at SMC are inferred to peak at 2 g in the damage zone (using another nearby station yielded a similar result). The low frequency variation, which was composed of 0.5 Hz energy arriving after the high frequency burst, remained at about the same amplitude. Although non-linear effects may invalidate such a linear extrapolation, this analysis supports the view that the high frequency pulses of acceleration were responsible for the focused damage, rather than the low frequency variation that followed. Single and double story structures are thought to be most susceptible to waves in the 5-10 Hz range. Thus even though the low frequency variation continues for tens of seconds after the main event, we believe this was not selectively focused, but may have contributed to exacerbating damage caused by the initial high frequency burst.

In a similar manner, we removed focusing effects to obtain the strong ground motion in the surrounding non-damaged zone by multiplying the SMC record with the transfer function between stations S95 and S56. The result shown in the lower panel of figure 2 has a peak acceleration of 0.4 g, similar to observations at nearby strong motion stations in sign)ficantly lessdamaged areas such as that at UCLA. Thus removal of focusing effects brings the SMC observations into agreement with calculations of synthetic seismograms by Zeng and Anderson
(1996) who found that SMC amplitudes were sign)ficantly greater than their average model fit to 12 strong motion stations.

Liu Hong continues work on her Ph.D. which has included finite-difference modeling of the phases in the Santa Monica aftershock data as well as documenting azimuthal amplification effects in the Sherman Oaks data. She has successfully modeled the double phase seen in the Santa Monica data (Gao et al., 1996).

2. Site selection and planning/proposal writing for a Vibroseis/Explosion survey of Santa Monica damage zone.

Along with T. Henyey and R. Clayton, we have proposed to SCEC, the USGS, and NSF to carry out LARSE II which includes a component to use active seismic methods to delineate the purported focusing structures beneath Santa Monica. Two 30 km auxiliary lines will be added to the main refraction-reflection line. One crosses the Santa Monica damage zone. The other, for comparison, is located to the east in the relatively undamaged zone. We propose to deploy 150 seismometers along these lines with shots and/or vibroseis profiles along the same lines. This summer in conjunction with the USGS refraction team led by Gary Fuis we traversed the lines and selected shot points.

3. Installation of 18 SCEC stations ~ 3 km spacing across LA basin to search for amplification/focusing effects in local, regional and teleseismic data.

While this experiment is primarily associated with group D activities, data from it will be of interest to group B scientists. To date, 9 months of recording have been completed. We have observed high amplitudes in the deepest part of the LA basin and low amplitudes in the Puente hills where waveforms become emergent suggesting a frequency-dependent transfer function, possibly explained by defocusing. This data base will provide an important resource for understanding basin amplification effects.

Figure 1. Map showing distribution of red-tagged buildings (squares), major highways (thin solid
lines), southern boundary of the SFV, and amplification factors derived from S waves (circles).
Figure 2. Stereographic projection of amplification factors in Sherman Oaks for S waves, relative
to the mean values for the stations on the Santa Monica Mountains. Relative amplitudes are given
by the size of the circles. Events from the north display much larger amplification factors than
those from the NW.
Figure 3. Same as Figure 2 but for P waves.
Figure 4. Velocity model used for finite difference waveform modeling for Sherman
Oaks.
Figure 5. SH waves for Sherman Oaks velocity structure showing focusing above the sub-basin.

Publications based on SCEC funded research

Davis, P. M., and L. Knopoff, The elastic modulus of media containing strongly interacting antiplane cracks, J. Geophys. Res., 100, 18,253-18,258, 1995.

Davis, P.M., and L. Knopoff, Reply, J. Geophys. Res., 101, 25,377-25379, 1996.
Gao, S., H. Liu, P.M. Davis, and L. Knopoff, Localized Amplification of Seismic Waves and Correlation with Damage due to the Northridge Earthquake, BSSA. Vol. 86, No. 1B, pp. S209-S230, Feb. 1996.

Fuis, G.S., D.A. Okaya, R.W. Clayton, W.J. Lutter, T. Ryberg, T.M. Brocher, T.M. Henyey, M.L. Benthien, P.M. Davis, J. Mori, R.D. Catchings, U.S. ten Brink, M.

Kohler, K.D. Klitgord, and R. Bohannon, Images of crust beneath southern California will aid

o study of earthquakes and their effects, EOS, Trans. Amer. Geophys. Union, 77, 18, 173,176, 1996.

Gao, S., H. Liu and P.M. Davis, A 98-station Seismic Array to Record Aftershocks of the 1994 Northridge Earthquake, USGS Open-File Report, 96-690, 1996.

Kohler, M. D., P. M. Davis, H. Liu, M. Benthien, S. Gao, G. S. Fuis, R. W. Clayton, D. Okaya, and J. Mori, Data Report for the 1993 Los Angeles Region Seismic Experiment (LARSE93), Southern California: a passive study from Seal Beach northeastward through the Mojave Desert, U.S. Geological Survey Open-File Report, 96-85, 82 pp., 1996.

Kohler, M. D., and P. M. Davis, Crustal thickness variations in Southern California from Los Angeles Region Seismic Experiment (LAPSE) passive phase teleseismic travel times, BSSA, 87,5, 1330-1344, 1997.

Kohler, M.D., J.E. Vidale, and P.M. Davis, Complex scattering within D" observed on the very dense Los Angeles Region Seismic Experiment passive array, Geophys. Res. Letters, 24,15,1855,1859, 1997.