S53C-08
The 1995 Mw 7.2 Gulf of Aqaba Earthquake revisited: Identifying active fault segments by joint inversion of geodetic and teleseismic data

Friday, 18 December 2015: 15:25
305 (Moscone South)
Hannes Bathke1, Guangcai Feng2, Sebastian Heimann3, Sigurjon Jonsson1 and Paul Martin Mai1, (1)King Abdullah University of Science and Technology, Thuwal, Saudi Arabia, (2)Central South University, Changsha, China, (3)Helmholtz Centre Potsdam GFZ German Research Centre for Geosciences, Potsdam, Germany
Abstract:
The largest earthquakes in Saudi Arabia occur at the northwestern boundary of the Arabian plate on a system of left-lateral transform faults extending from the Red Sea in the South and North through the Gulf of Aqaba.

The last major earthquake along this boundary occurred in November 1995 and in a complex tectonic setting offshore in the Gulf of Aqaba, consisting of several transform faults and pull-apart basins.

Various authors have studied this earthquake in the past, either by using geodetic radar (InSAR) or teleseismic (P and S waves) data, and several source models of the earthquake rupture and the active fault segments have been proposed. However, these source models differ significantly from each other and it still remains unclear which fault segments within the Gulf were activated during the event. There are various reasons for these differences. Teleseismic data alone cannot locate the event well, whereas the lack of near field co-seismic displacement data (due to the event’s offshore location) and the quasi north-south oriented strike-slip faulting of the earthquake result in a low SNR in the radar data. Consequently, the uncertainties of inferred model parameters are large and have not been properly estimated so far.

In this work, we use radar data from two additional tracks that have not been used before, which provides a more complete displacement field of the earthquake. By using multiple aperture radar interferometry it is possible to better constrain the south-north oriented strike-slip component. In addition, we include both the geodetic data and the teleseismic data in a joint inversion setup allowing combining the strengths of each dataset to constrain the model parameters. By including the full data-variance covariance-matrixes in Bayesian inference sampling, we estimate the model-uncertainties and the related range of likely source models. Consequently, we re-evaluate, which fault segments were activated during the earthquake in the Gulf of Aqaba.

Our results show that the InSAR data cannot alone properly constrain the parameters of the activated faults within the gulf. In addition, the teleseismic data clearly show that the earthquake consisted of a few sub-events and that the earthquake ruptured from South to North close to the eastern shore of the Gulf of Aqaba.