Application of Acoustic FWI in a Deep-Water Environment Using Towed Streamer Data
M.A.A.M. Jelani, R.G. Pratt and A.D. Booth
Event name: 80th EAGE Conference and Exhibition 2018
Session: Poster: FWI C
Publication date: 11 June 2018
Info: Extended abstract, PDF ( 16.09Mb )
Price: € 20
FWI to extract velocity models has been successfully applied in shallow-water environments by exploiting refracted and turning wave information. In deep-water environments with a limited recorded offset length (~10 km), FWI is more challenging due to there being less refracted energy to control the inversion process. We demonstrate the applicability of acoustic FWI method in a deep-water environment with towed streamer seismic data, acquired in the Orange Basin, South Africa. To do this, we use a highly accurate traveltime tomographic image (Jelani and Booth, 2017) as a starting model, and we initially invert the very lowest frequency components in order to mitigate the non-linearity issue. We applied a variety of strategies of data pre-processing, pre-conditioning, wavenumber filtering, and re-estimating the source signatures to optimize the FWI results. The inverted velocity model was validated by a comparison of the observed and predicted waveforms. The kinematics of the predicted wavefield shows a reasonably good agreement with the observed waveform as well as a good correlation with the velocity image corresponding to the reflector on the PSDM seismic image. Further advances are anticipated to incorporate amplitude information with higher frequencies >6.0 Hz, in order to investigate the attenuation effects within the modelled space.