 |
| Titel |
Sequence-based Parameter Estimation for an Epidemiological Temporal Aftershock Forecasting Model using Markov Chain Monte Carlo Simulation |
| VerfasserIn |
Fatemeh Jalayer, Hossein Ebrahimian |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250097951
|
| Publikation (Nr.) |
 EGU/EGU2014-13580.pdf |
|
|
|
|
|
| Zusammenfassung |
| Introduction
The first few days elapsed after the occurrence of a strong earthquake and in the presence
of an ongoing aftershock sequence are quite critical for emergency decision-making
purposes. Epidemic Type Aftershock Sequence (ETAS) models are used frequently for
forecasting the spatio-temporal evolution of seismicity in the short-term (Ogata, 1988).
The ETAS models are epidemic stochastic point process models in which every
earthquake is a potential triggering event for subsequent earthquakes. The ETAS
model parameters are usually calibrated a priori and based on a set of events that do
not belong to the on-going seismic sequence (Marzocchi and Lombardi 2009).
However, adaptive model parameter estimation, based on the events in the on-going
sequence, may have several advantages such as, tuning the model to the specific
sequence characteristics, and capturing possible variations in time of the model
parameters.
Simulation-based methods can be employed in order to provide a robust estimate for the
spatio-temporal seismicity forecasts in a prescribed forecasting time interval (i.e., a day)
within a post-main shock environment. This robust estimate takes into account the
uncertainty in the model parameters expressed as the posterior joint probability
distribution for the model parameters conditioned on the events that have already
occurred (i.e., before the beginning of the forecasting interval) in the on-going seismic
sequence. The Markov Chain Monte Carlo simulation scheme is used herein in
order to sample directly from the posterior probability distribution for ETAS model
parameters. Moreover, the sequence of events that is going to occur during the
forecasting interval (and hence affecting the seismicity in an epidemic type model like
ETAS) is also generated through a stochastic procedure. The procedure leads to two
spatio-temporal outcomes: (1) the probability distribution for the forecasted number of
events, and (2) the uncertainty in estimating the probability of exceeding a certain
number of events, both with a magnitude greater than a prescribed threshold. The
robust ETAS forecasts can be directly implemented in adaptive daily aftershock
hazard and risk assessment procedures (see Jalayer et al. 2011, Ebrahimian et al.
2014).
Numerical example
As the numerical example, the L’Aquila 2009 (central Italy) aftershock sequence is used
herein. The methodology described in the previous section is applied in order to perform
robust forecasting for the spatio-temporal distribution of number of events within the first few
days elapsed after the main-shock.
References
Ebrahimian, H., Jalayer, F., Asprone, D., Lombardi, A.M., Marzocchi, W., Prota, A.,
Manfredi, G. (2014). “Adaptive daily forecasting of seismic aftershock hazard.” Bull.
Seismol. Soc. Am., DOI:10.1785/0120130040.
Jalayer, F., Asprone, D., Prota, A., Manfredi, G. (2011) “A decision support system for
post-earthquake reliability assessment of structures subjected to aftershocks: an application to
L’Aquila earthquake, 2009.” Bull. Earthq. Eng., 9(4), 997-1014.
Marzocchi, W., and Lombardi, A.M. (2009). “Real-time forecasting following a
damaging earthquake.” Geophys. Res. Lett. 36, L21302, doi:10.1029 / 2009GL040233.
Ogata, Y. (1988). “Statistical models for earthquake occurrences and residual analysis for
point processes.” J. Am. Stat. Assoc., 83, 9-27. |
|
|
|
|
|
|