Stochastic salt modeling for characterizing seismic imaging uncertainties.

in: 2019 Ring Meeting, ASGA

Abstract

Seismic imaging of salt environments is prone to large uncertainties due to the strong velocity variations introduced by the presence of salt bodies. To investigate these uncertainties, we present early results on the characterization of the salt variability impact on the quality of seismic images. We use a recent method to automatically generate stochastic realizations of salt bodies that we extend for modeling salt welds. The modeling of salt welds consists in the segmentation of an implicit scalar field representing the salt envelope, combined with several visibility criteria that define the actual extension of the weld. The next step consists in building velocity models from these structural models. We designed a method that allows for the semiautomatic generation of velocity models from stochastically generated salt models. To handle the potential presence of welds, it relies on the combination of several predefined background sediment velocity models, on top of which is overlaid the actual salt body velocity. As we focus on the impact of salt variability, we assume the exact background velocity model in the sediments around salt is known. To investigate the uncertainties underlying seismic imaging, we use one of the generated velocity models to simulate a seismic acquisition. We then perform a reverse-time migration for each velocity model to obtain a set of alternative seismic images. The comparison of the different images provides first insights on the impact of salt misinterpretation on seismic images, and paves the way to stochastic velocity model updating.

Download / Links

BibTeX Reference

@inproceedings{ClausollesRM2019,
 abstract = { Seismic imaging of salt environments is prone to large uncertainties due to the strong velocity variations introduced by the presence of salt bodies. To investigate these uncertainties, we present early results on the characterization of the salt variability impact on the quality of seismic images. We use a recent method to automatically generate stochastic realizations of salt bodies that we extend for modeling salt welds. The modeling of salt welds consists in the segmentation of an implicit scalar field representing the salt envelope, combined with several visibility criteria that define the actual extension of the weld. The next step consists in building velocity models from these structural models. We designed a method that allows for the semiautomatic generation of velocity models from stochastically generated salt models. To handle the potential presence of welds, it relies on the combination of several predefined background sediment velocity models, on top of which is overlaid the actual salt body velocity. As we focus on the impact of salt variability, we assume the exact background velocity model in the sediments around salt is known. To investigate the uncertainties underlying seismic imaging, we use one of the generated velocity models to simulate a seismic acquisition. We then perform a reverse-time migration for each velocity model to obtain a set of alternative seismic images. The comparison of the different images provides first insights on the impact of salt misinterpretation on seismic images, and paves the way to stochastic velocity model updating. },
 author = { Clausolles, Nicolas AND Collon, Pauline AND Caumon, Guillaume AND Irakarama, Modeste },
 booktitle = { 2019 Ring Meeting },
 publisher = { ASGA },
 title = { Stochastic salt modeling for characterizing seismic imaging uncertainties. },
 year = { 2019 }
}