HR: 08:45h
AN: V31E-04    [Abstracts]
TI: Bayesian and Geostatistical Methodologies For The Long-Term Estimation of Volcanic Hazards
AU: Jaquet, O
EM: olivier.jaquet@colenco.ch
AF: Colenco Power Engineering Ltd, Taefern 26, Baden, 5405 Switzerland
AU: * Connor, C
EM: cconnor@chuma1.cas.usf.edu
AF: University of South Florida, Department of Geology 4202 E. Fowler Ave. , Tampa, FL 33620 United States
AU: Connor, L
V31E-04 AF: University of South Florida, Department of Geology 4202 E. Fowler Ave. , Tampa, FL 33620 United States
AB: Due to the difficulty of describing the complex spatial and temporal patterns inherent to volcanism, the use of solely deterministic models is not sufficient for long term estimation of volcanic hazards. In order to account of the intrinsic uncertainty of volcanism that occurs in space, time and with respect to event types and their intensity, the use of probabilistic models becomes quite natural for long term hazard assessment. The motivation of this on-going work is driven by the necessity of investigating various probabilistic approaches for the estimation of volcanic hazards in relation to high-level radioactive waste repository siting in Japan. In particular, the integration of additional sources of uncertainties linked to the distribution of volcanic events in space as well as to their occurrence need to be accounted for by the modelling. The concepts of the Bayesian and geostatistical models of the proposed approach will be presented and then illustrated by a case study using data from the Tohoku volcanic arc, with particular emphasis on modelling future distribution and character of volcanism, given the geophysical setting of the volcanic arc, including the relationship between volcano distribution, isostatic and Bouguer gravity anomalies and topography, which are consistent with magma accumulation at the Moho and Conrad discontinuities.
DE: 8488 Volcanic hazards and risks
SC: Volcanology, Geochemistry, Petrology [V]
MN: Fall Meeting 2005