 |
| Titel |
Cooperative inversion of magnetotelluric and seismic data sets |
| VerfasserIn |
M. Markovic, F. Santos |
| Konferenz |
EGU General Assembly 2012
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 14 (2012) |
| Datensatznummer |
250060088
|
|
|
|
|
|
| Zusammenfassung |
| Cooperative inversion of magnetotelluric and seismic data sets
Milenko Markovic,Fernando Monteiro Santos
IDL, Faculdade de Ciências da Universidade de Lisboa
1749-016 Lisboa
Inversion of single geophysical data has well-known limitations due to the non-linearity of the fields
and non-uniqueness of the model. There is growing need, both in academy and industry to use two or
more different data sets and thus obtain subsurface property distribution. In our case ,we are dealing
with magnetotelluric and seismic data sets.
In our approach,we are developing algorithm based on fuzzy-c means clustering technique, for pattern
recognition of geophysical data. Separate inversion is performed on every step, information exchanged
for model integration. Interrelationships between parameters from different models is not required in
analytical form. We are investigating how different number of clusters, affects zonation and spatial
distribution of parameters. In our study optimization in fuzzy c-means clustering (for magnetotelluric
and seismic data) is compared for two cases, firstly alternating optimization and then hybrid method
(alternating optimization+ Quasi-Newton method).
Acknowledgment: This work is supported by FCT Portugal |
|
|
|
|
|
|