Probabilistic techniques for estimating earthquake triggered landslide hazards

  • Joseph M. Mankelow

    Student thesis: Doctoral Thesis

    Abstract

    The Summit Ridge area, situated in the North Santa Cruz Mountains,
    California, was heavily affected by landsliding during the 1989 Loma Prieta
    earthquake( MW 7.0). Widespread shallow, along with 17 deep-seated,la ndslides
    caused extensive damage.
    In order to study possible techniques for assessing landslide hazard a geographical information system (GIS) was utilised. Geological, geotechnical, geomorphological and seismological data were utilised during deterministic analyses using two standard earthquake slope stability models (pseudo-static and Newmark displacement). The models were used to assess the landslide potential that existed during the Loma Prieta event. It was found that, of these two models, the Newmark displacement method proved most successful at predicting the location of shallow unstable slopes. However, because both models are deterministic in nature, they did, not take into account any of the error and uncertainty in the input parameters. Therefore, a probabilistic Newmark displacement analysis technique was developed. The new technique allowed the estimation of the probability that a slope will exceed a certain critical value of Newmark displacement. Use of such an approach resulted in a more realistic distribution of hazard when compared with the distribution of actual landslides triggered by the 1989 earthquake.
    As a result of this seismic hazard assessment, the new probabilistic technique
    was used to undertake a shallow landslide hazard assessment for a design earthquake located on the Northern East Bay Segment of the Hayward fault. Results have indicated that during the next 30 years many slopes have greater than a 20 % chance of failure.
    In an attempt to assess the contribution modem simulation techniques could
    make to assessing earthquake triggered landslides the data obtained from one of the deep-seated landslides triggered during the, Loma Prieta earthquake was input into a recently developed simulation application. Input probability distributions were sampled in an iterative manner and the resulting values were input into the empirical relationship for estimating Newmark displacement. Ibis resulted in the production of an associated output distribution which was used to derive the probabilities associated with increasing displacement categories. The technique utilised proved very successful and charts depicting the probabilities for exceeding increasing Newmark displacements on the deep-seated landslide during potential earthquakes located on four of the principal fault segments in the San Francisco Bay Area have been produced.
    Date of Award1998
    Original languageEnglish
    SupervisorWilliam Murphy (Supervisor) & Dave Giles (Supervisor)

    Cite this

    '