Use this URL to cite or link to this record in EThOS:
Title: On the use of micro-earthquakes to predict strong ground-motion
Author: Edwards, Benjamin
Awarding Body: University of Liverpool
Current Institution: University of Liverpool
Date of Award: 2008
Availability of Full Text:
Access from EThOS:
Access from Institution:
Since the prediction of earthquakes is, at least at present. not possible, the understanding of energy propagation from an earthquake is important in terms of mitigating structural and human losses. The physical manifestation of an earthquake is ground shaking and it is this that ultimately causes structural failure, and in the worst case scenario. loss of life. A commonly used measure of damage likelihood to a particular structure is the peak ground acceleration or velocity (PGA or PGV respectively). Typically. the prediction of PGA and PGV from a particular earthquake scenario is performed through the regression of recordings of peak ground motion. This provides predictive relations. or attenuation relations. that are valid over the magnitude range defined by the data. Unfortunately. in regions of low or moderate seismicity, the availability of recordings of strong ground-motion is unlikely. In this case the attenuation relations that are typically used have been derived from similar tectonic regions. but with higher seismicity. In doing this a bold assumption must be made: that the scaling of seismic energy is not regionally dependent. In this thesis an alternative approach is taken, ''- where this assumption is not required. Instead, commonly available micro-earthquake data are utilised. From this data the attenuating parameters of the crust are derived along with scaling properties of the earthquake source. In order to obtain and robustly resolve a notoriously non-unique solution, a tomographic reconstruction of Q and K is adopted. This enables the stable decoupling of geometrical decay, seismic moments. and the Brune stress drop. Using these crustal attenuation and source scaling parameters. a stochastic method is used to predict peak ground motion in terms of the 5% of critical damped response spectrum. The method is carefully tested in terms of its ability to reconstruct a synthetic dataset of known input parameters. Additionally we test the effect of varying the synthetic and inversion models in order to analyse parameter trade-off. Synthetic data are produced using a variety of methods: forward modelling of the inversion model. finite difference fault modelling and stochastic simulation. In addition bootstrap analyses are performed to estimate errors on the resultant models. It is found that the model parameters are strongly covariate. but it is shown that through the use of the tomographic reconstruction of Q, a robust solution is obtained. In addition this allows the use of a multi-dimensional Q structure as opposed to the typical homogeneous model found in other studies. The method was applied to a seismically active region: central Japan. It was shown that the results obtained using only micro-earthquake data were comparable to those obtained using strong-motion data from the same area. Additionally the response spectra for large earthquakes in Japan were modelled using a stochastic method along with the attenuation parameters derived using micro-seismicity. This was shown to be successful by comparing the model to empirical data from large earthquakes that occurred in the study region. With the knowledge that the methodology used was successful in the case of Japan it was applied to UK microearthquake data. Predictive ground motion relations were derived and shown to be valid over the magnitUde range on record. The model was compared to other predictive relations used previously in the UK. It was shown that these alternative relations significantly over-predicted PGA and PGV for small earthquakes in the UK. Additionally the model was used to successfully predict ground motion for two of the largest UK earthquakes in the last three decades.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral