HR: 1330h
AN: OS22B-1157    [PDF]
TI: A Framework for the Probabilistic Analysis of Tsunami Hazards
AU: * Geist, E L
EM: egeist@usgs.gov
AF: U.S. Geological Survey, 345 Middlefield Rd., MS 999, Menlo Park, CA 94025 United States
AB: A framework for analysis of tsunami hazards is proposed that is based on probabilistic seismic hazard analysis (PSHA), though with some important modifications. PSHA is a convenient methodology to adopt for the analysis of tsunami hazards in part because, for most regions in the United States, seismic source model specifications have been developed through the U.S. Geological Survey's National Seismic Hazard Mapping Project. The two main modifications to PSHA that are needed are (1) incorporation of far-field sources (i.e., distant earthquakes) that can generate damaging tsunamis and (2) a numerical hazard model (i.e., runup and inundation in tsunami hazard analysis) rather than an empirical model (i.e., ground motion in seismic hazard analysis). An important step in the development of probabilistic tsunami hazard analysis (PTHA) is the definition of uncertainties. For local tsunamis, variation in slip distribution patterns is a significant source of uncertainty that is inherently unpredictable (i.e., an aleatory uncertainty). Stochastic source models can be used in PTHA to estimate the range of possible nearshore tsunami amplitudes for a given magnitude and rupture geometry. Other important sources of uncertainty for local tsunami generated by subduction zone earthquakes include occurrence of secondary faulting near the trench, and triggered submarine and coastal landslides. For distant tsunamis, in addition to the fundamental tsunami source parameters of seismic moment and epicentral distance, source radiation pattern and propagation raypath are important factors in hazard calculations. Tsunami propagation from both local and distant sources is also sensitive to nearshore variations in bathymetry (which itself maybe a source of epistemic uncertainty) that is analogous to site response in PSHA. In contrast to past method for forecasting tsunami hazards, a PSHA-based method permits incorporation of uncertainties into the probabilistic forecast. PTHA as outlined in this study is compared to log-frequency/maximum runup height curves derived from historical data that were used in the past, especially in light of new probability analyses of complex systems. An important final consideration for PTHA is how probabilistic results are formatted and presented to different end-user communities.
DE: 3210 Modeling
DE: 4564 Tsunamis and storm surges
DE: 6309 Decision making under uncertainty
SC: Ocean Sciences [OS]
MN: 2003 Fall Meeting