CONTACT: Stanford University News Service (415) 723-2558
Quake collapsed roadway where ground was softest
STANFORD -- The ground shook 1.5 times more intensely under a section of elevated roadway that collapsed in the 1989 Loma Prieta earthquake, killing 42 people, than at the other end of the viaduct anchored in firmer soil, Stanford researchers say.
Their quantification of ground shaking under the Cypress viaduct in Oakland, Calif., provides further evidence that earthquake damage patterns can be predicted to prevent disasters such as that during the afternoon rush hour on Oct. 17, 1989.
"Property owners, government planners and insurers should be using soil information in earthquake-prone areas to guide their decisions about the construction of roads, hospitals and other critical facilities," said Haresh Shah, chairman of Stanford's civil engineering department and one of the researchers who conducted the study.
Shah and Keiichi Tamura, a visiting researcher at Stanford until recently, studied ground motion at 300 points along the 11-block-long elevated roadway in Oakland. The north portion of the structure, which collapsed, was built on artificial fill over the top of soft San Francisco Bay mud. The south portion of the viaduct, which did not collapse, was founded on a firm Merritt sand stratum.
They found a ground-motion spectral intensity 1.5 times greater in the mud than in the sand. Spectral intensity takes into account the frequency and the energy of ground motion.
Other measurements of ground shaking, such as the maximum acceleration during the earthquake and maximum ground displacement, were also greater on the end of viaduct collapse, but the spectral intensity of the shaking proved to be the best predictor of damage to the structure, according to Shah and Tamura, who is now a researcher for the Ministry of Construction in Tsukuba, Japan.
No ground-motion measurement devices were actually present at the site of the viaduct during the earthquake. To do their analysis, the researchers began with measurements taken on Bay mud in the town of Emeryville, nine-tenths of a mile north of the collapsed portion of the viaduct. They used the north-south component of the recorded ground motion, rather than the east-west component to calculate characteristics of the earthquake at 300 points along the viaduct.
The east-west motion of the Loma Prieta earthquake was greater than the north-south motion, but the researchers used the north-south data, because the viaduct extends in an almost north- south direction and because the type of soil also changes underneath it in the north-south direction.
They found that maximum acceleration on the ground surface was greatest at points overlying the mud, and the change in acceleration was steepest at points at the border between the mud and the sand.
The maximum displacement of ground varied from 3 centimeters at one end of the viaduct to 3.7 centimeters at the other, with the peak of 4.2 centimeters at the point where the Bay mud is deposited deepest, which is where the viaduct collapsed.
The intensity of the shaking varied less along the structure's length than the other parameters measured, but the spectral intensity difference was the most marked difference between the soft soil deposits where the collapse occurred and the firm soil deposits where it did not, Shah said.
The results demonstrate that it is possible with current knowledge to estimate the intensity of ground shaking at a site by taking into account soil conditions, distance from the epicenter and size of future earthquake events, Shah said.
"Land-use planning and building design strategies can be developed to minimize damage, based on our current ability to estimate ground-motion intensity," Shah said.
Stanford, for instance, is developing a seismic-risk mitigation plan for strengthening existing campus facilities. A critical element in developing the design guidelines will be a ground-motion analysis for the 8,200-acre campus. (See separate press release.)
An artificial-intelligence-based system developed at Stanford's John A. Blume Earthquake Engineering Center has the soil data base to estimate ground motion intensity and the performance of various types of buildings for any address in California, Shah said. Similar systems are under development by others.
Risk Management Software Inc., Mountain View, Calif., has a license from Stanford to market the Stanford-developed Information Risk Assessment System. It has been used by insurance companies and real-property investors.
This is an archived release.
This release is not available in any other form.
Images mentioned in this release are not available online.