Estimation parameters of fault causing earthquake using co-seismic displacement field and artificial neural networks Case study: The 2005 Qeshm Island earthquake (Iran)

The ultimate goal of geophysical observations is determining geological structures from the geophysical data which is very difficult due to the complex structure of the earth’s interior. Retrieving the parameters of a fault through a set of its deformation observations is called “geophysics’ inverse problem”. In this article، the Artificial Neural Network as an evolutionary method are applied to retrieve the Qeshm fault parameters. In this research، in order to estimate parameters of the fault causing the earthquake، Insar displacement observations of the earthquake occurred on The 2005 November 27 in Qeshm and also neural network have been used. Deploying this approach in co-seismic displacement field of Qeshm leaded to parameter estimation of the fault with a depth، length، widths، dip angle، strike، slip، of 6. 1 km، 7. 4 km، 4. 6 km، 42 degrees، 251 degrees and 88 cm، respectively. Maximum difference of the reconstructed displacement field based on the estimated parameters on displacement field obtained by Insar is demonstrated as large as 1 cm in vertical direction and 4 cm in north-south direction. The results indicate that the artificial neural networks are trusted approaches in parameter estimation of faults.