Development and Tuning of a 3-D Stochastic Inversion Methodology for the European Arctic
Abstract
High-resolution seismic models are a critical component of calibrating earth structure for improved seismic monitoring. We will in this study develop the Markov Chain Monte Carlo (MCMC) inversion method into an even stronger tool for deriving reliable three-dimensional seismic models of the crust and upper mantle, based on multiple types of geophysical data sets. This will be done by tuning the method to the European Arctic through development of a probabilistic geophysical model. While a new and much improved model (BARENTS3D) recently has been developed for this region (Ritzmann et al., 2007), stochastic models have a potential to better represent our state of knowledge (and uncertainty) about geophysical structure because deterministic models do not express well the tradeoffs inherent in the data. Stochastic inverse methods also allow a more systematic exploration of the model space to help avoid the trap of falling into local minima. Finally, stochastic models allow prediction of observable distributions (and through them observable uncertainties).
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 2008
- Accession Number
- ADA516012
Entities
People
- Hilmar Bungum
- Jan I. Faleide
- Michael E. Pasyanos
- Stephen A. Clark
Organizations
- NORSAR