Loading...
Uncertainty quantification in seismic imaging
Pawelec, Iga
Pawelec, Iga
Citations
Altmetric:
Advisor
Editor
Date
Date Issued
2018
Date Submitted
Collections
Research Projects
Organizational Units
Journal Issue
Embargo Expires
Abstract
To make informed decisions, one has to consider all available knowledge about the assessed problem. An important part of the decision-making process is understanding uncertainties and how they influence the outcome. In seismic exploration, many decisions are based on interpretations of seismic images, which are affected by multiple sources of uncertainty. Thus, image uncertainty quantification is an important, albeit challenging task. In this thesis, I focus on two uncertainty sources that affect seismic imaging: data uncertainty and velocity uncertainty. I quantify the seismic data uncertainty using theoretical analysis applied to two field experiments with repeated shots. My analysis reveals that amplitude distributions for each data sample as a function of time and position are not Gaussian and that the uncertainty of a seismic event is proportional to its mean amplitude. I also find that seismic events excited by the source are highly repeatable, but small changes of the source position impact the amplitude response, highlighting the importance of geometry repeatability for the lapse studies. Velocity uncertainty also has a large impact on image uncertainty, as it affects reflector positioning and the focusing of seismic events. By examining two subsalt imaging scenarios with geological uncertainty caused by the salt body physical properties, I demonstrate that image uncertainty, expressed as a function of the image amplitude or as a function of the reflector location, is the largest under the salt: the image amplitude distributions are two times broader under the salt than away from it. The confidence index maps are a useful tool to convey the information about image amplitude uncertainty to an interpreter, while the location uncertainty reveals uncertain directions and is affected by acquisition geometry. The main challenges facing uncertainty quantification in seismic imaging include integration of different sources of uncertainty and reducing the computational cost of the analysis. My analysis leads to recommendations about possible approaches towards these challenges, with emphasis on using sparsity to reduce the dimensionality of the problem.
Associated Publications
Rights
Copyright of the original work is retained by the author.