 |
| Titel |
Estimating the uncertainty of seismic point source solutions |
| VerfasserIn |
Simon C. Stähler, Kasra Hosseini, Ran Zhang, Karin Sigloch |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250094498
|
| Publikation (Nr.) |
 EGU/EGU2014-9911.pdf |
|
|
|
|
|
| Zusammenfassung |
| A point source is the most basic description of an earthquake. While it is obviously
too simplistic for large earthquakes (-¿ M7.5), it is still sufficient for intermediate
(M5.5 - M7.5) earthquakes, especially in sparsely instrumented areas and when large
numbers of earthquakes are to be processed automatically. Seismic tomography
regularly needs a large number of point source solutions to infer the structure of the
Earth.
We present a Bayesian inference of seismic point source parameters, including depth, full
moment tensor and the source time function.
A focus of this study is the correct handling of noise in data and modelling error in the
forward calculation of seismic waveforms. We show that samplewise misfits like the -1 or -2
norm are not suited in presence of strong forward modelling errors and instead propose to use
the cross-correlation coefficient CC as a misfit criterion. We further derive a likelihood
function for CC to allow for full Bayesian inference. This includes an estimation of the
correlation of measurements at different stations.
The result of this study can be used as an novel input for seismic tomography, especially
since it allows to estimate the uncertainty of input parameters for tomography, like travel-time
or amplitude anomalies. |
|
|
|
|
|
|