Probabilistic Procedures for Peak Ground Motions

by John A. Blume, (Hon.M.ASCE), President; URS/John A. Blume & Associates, Engrs., San Francisco, Calif.,
Anne S. Kiremidjian, (M.ASCE), Asst. Prof.; Dept. of Civ. Engrg., Stanford Univ., Stanford, Calif.; formerly, Research Assoc., URS/John A. Blume & Associates, Engrs., San Francisco, Calif.,

Serial Information: Journal of the Structural Division, 1979, Vol. 105, Issue 11, Pg. 2293-2311

Document Type: Journal Paper


Three methods of estimating the frequency of earthquaake occurence are presented: regression analysis of past earthquake data; integration of fault dislocation data over long time periods; and consideration of plate boundary dislocation rates. The models consider fault rupture as a function of magnitude when faults can be modeled as line sources. When locations and activity rates are not known, seismically active regions are modeled as diffused area sources. The equation for attenuation of ground motion includes a soil impedance factor and a lognormal probability distribution on the data scatter. By applying the three procedures to a site for a nuclear power plant in southern California, it was determined that the plate boundary method yields lower peak ground motion values than the other methods at the same risk level. The highest peak ground acceleration values were predicted by the long-period (20,000,000 yr) fault-dislocation method.

Subject Headings: Ground motion | Geological faults | Probability | Regression analysis | Displacement (mechanics) | Plates | Domain boundary | Power plants | California | United States

Services: Buy this book/Buy this article


Return to search