A survey of the seismic hazards for about 30,000 nonresidential buildings in Portland, Oregon was conducted, and an earthquake damage and loss estimation model was developed. To conduct the buildings survey, ATC-21's Rapid Screening Procedure was used. For each building the data included: address, year built, area, number of stories, building name, use, occupancy type, estimated range of number of people, nonstructural falling hazards, building type, and performance modifiers such as soft story and torsion. Soil data were provided by the Earthquake Hazard Maps published by the Oregon Department of Geology and Mineral Industries. The algorithm set up on the GIS was to compute the structural score based on ATC-21 score, modify the score based on the geologic condition of the site, and finally, compute the damage. The proposed loss estimation model is based on ATC-13, and the scoring system developed by ATC-21. The results are in terms of percent (or dollar) damage to buildings in the survey and loss of life and serious injuries.

Recent geologic research has shown that earthquakes more destructive than formerly expected are likely to occur in the Pacific Northwest. To mitigate catastrophic loss, planners are gathering information to make decisions on implementing regional seismic retrofit programs.This research develops a model to estimate regional earthquake losses for existing buildings, and determine optimal retrofit priorities and budgets.Fragility curves are developed to provide earthquake damage estimates for a range of seismic intensities. The published earthquake damage estimates of a large group of prominent earthquake engineering experts are extended to include the combined effect of structure type, earthquakesensitive variations in building design, site-specific soil conditions, and local seismic design practice.

LocK Buttress Diaphragm Wal 1 Uti I izing Pressuremeter Test Results. APPROVED BY MEMBERS OF THE THESIS COMMITTEE:M.M. Chairman . rvif'lH~esonThe P-y curve, used in current practice as an efficient I ine-load vs. soi I displacement model for input into the finite difference method of laterally loaded pi le ana I ys is, is extended in this study for use with The results indicate that preloading wi 11 reduce wall deflections by at least 4-fold, but that wal I cracKing can potential Jy double deflections. Safety factors against passive soi I fai Jure were determined to be about 5 at anchor pre load, and more than 40 after ful I excavation.