Multi-scenario interpretations from sparse fault evidence using graph theory and geological rules

in: 2018 Ring Meeting, ASGA

Abstract

Structural uncertainties arise during the interpretation of sparse and ambiguous fault evidence (limited outcrop observations, two-dimensional seismic lines, or focal mechanisms). Interpreters use the regional context and geological concepts to determine which of the disjoint fault evidence may actually belong to the same fault. We propose to represent each interpretation scenarios using a graph whose nodes are the fault evidence and edges are potential evidence associations. We describe a new stochastic algorithm that generates likely fault association interpretation scenarios. User-defined numerical rules aiming at representing the prior geological knowledge that constrains the simulation process. The algorithm is illustrated on a dataset made of synthetic seismic lines extracted from high-resolution seismic data, offshore Brazil. We show that this algorithm is able to generate a large range of possible scenarios in order to rigourously adress the data association problem in structural uncertainty studies. We led a sensitivity analysis on chosen geological rules that formally confirms that integrating geological knowledge into structural interpretation consistently reduces structural uncertainties. However, we show the difficulty of retrieving the reference association from the sparse data. Using restrictive and consistent rules increases the likelihood to recover this reference association.

Download / Links

BibTeX Reference

@inproceedings{RUNKJRM77,
 abstract = { Structural uncertainties arise during the interpretation of sparse and ambiguous fault evidence
(limited outcrop observations, two-dimensional seismic lines, or focal mechanisms). Interpreters
use the regional context and geological concepts to determine which of the disjoint fault evidence
may actually belong to the same fault. We propose to represent each interpretation scenarios using
a graph whose nodes are the fault evidence and edges are potential evidence associations. We
describe a new stochastic algorithm that generates likely fault association interpretation scenarios.
User-defined numerical rules aiming at representing the prior geological knowledge that constrains
the simulation process. The algorithm is illustrated on a dataset made of synthetic seismic lines
extracted from high-resolution seismic data, offshore Brazil. We show that this algorithm is able
to generate a large range of possible scenarios in order to rigourously adress the data association
problem in structural uncertainty studies. We led a sensitivity analysis on chosen geological rules
that formally confirms that integrating geological knowledge into structural interpretation consistently
reduces structural uncertainties. However, we show the difficulty of retrieving the reference
association from the sparse data. Using restrictive and consistent rules increases the likelihood to
recover this reference association. },
 author = { Godefroy, Gabriel AND Caumon, Guillaume AND Bonneau, Francois AND Laurent, Gautier },
 booktitle = { 2018 Ring Meeting },
 publisher = { ASGA },
 title = { Multi-scenario interpretations from sparse fault evidence using graph theory and geological rules },
 year = { 2018 }
}