SSA Print Logo Seismological Society of America
Skip to Content

Abstracts Online

A Bayesian Ground Motion Model for Estimating the Covariance Structure of Ground Motion Intensity Parameters

KUEHN, N.M., IEES, University of Potsdam, Potsdam/Germany, nico@geo.uni-potsdam.de; RIGGELSEN, C., IEES, University of Potsdam, Potsdam/Germany, riggelsen@geo.uni-potsdam.de; SCHERBAUM, F., IEES, University of Potsdam, Potsdam/Germany, fs@geo.uni-potsdam.de; ALLEN, T., Geoscience Australia, Canberra/Australia, Trevor.Allen@ga.gov.au

We present a Bayesian ground motion model that directly estimates both coefficients and the correlation between different ground motion intensity parameters. For this purpose, we set up a multivariate statistical model, embedded in a graphical framework, which mimics our insight into the data generating process, i.e. which includes a source, path and station term. For each term, coefficients to predict the median of the intensity parameter distribution can be estimated, together with the associated covariance structure (i.e. inter-event, intra-event and inter-station variability plus correlation coefficients). The graphical structure provides intuitive insight into the model. The coefficients of the model are estimated in a Bayesian framework using Markov Chain Monte Carlo simulation. Thus, prior information can be included in a principled way and an estimate of the epistemic uncertainty of the parameters can be obtained. The Bayesian approach also allows to update the model once new data is available. The parameters of the model are estimated on a global dataset using peak ground acceleration, peak ground velocity and the response spectrum at three periods as the target variables. The analysis shows that the coefficients of the model are similar to those estimated without the covariance structure. This is also true for the intra-event variability, while the inter-event variability is reduced when estimated with the covariance structure.

[Back]

Last Modified: 2011 Aug 10

Web: www.seismosoc.org