Appraising structural models using seismic data : misfit functions

in: 2017 RING Meeting, pages 1--17, ASGA

Abstract

Structural interpretation from a seismic image can be highly subjective, especially in complex geo- logical settings. A single seismic image will often allow multiple geologically valid interpretations. In this paper, we continue to investigate the problem of appraising structural interpretations using seismic data, with a strict focus on misfit functions. Misfit functions are a key component of struc- tural model appraisal. Understanding the relation between misfit functions and structural models is the first step towards building structural models that are optimally constrained by seismic data, and therefore less prone to interpretation uncertainties; it is also the first step towards automated interpretation quality control of structural models. We formalize the concept of appraising struc- tural models using seismic data. We start by giving a formal definition of what it means to appraise structural models using data; from this definition, we derive a set of conditions that are imposed on data space misfit functions for such an appraisal to be considered reliable. There are two major development in this paper. First, unlike in our previous studies, the conditions presented here no longer depend on some unknown reference model. Second, we show that there exist a special case of practical interest where all the conditions needed for appraising interpretations using vertical seismic profiling data can be satisfied

Download / Links

BibTeX Reference

@INPROCEEDINGS{Irakarama2017,
    author = { Irakarama, Modeste and Cupillard, Paul and Caumon, Guillaume and Sava, Paul and Edwards, Jonathan },
     title = { Appraising structural models using seismic data : misfit functions },
 booktitle = { 2017 RING Meeting },
      year = { 2017 },
     pages = { 1--17 },
 publisher = { ASGA },
  abstract = { Structural interpretation from a seismic image can be highly subjective, especially in complex geo- logical settings. A single seismic image will often allow multiple geologically valid interpretations. In this paper, we continue to investigate the problem of appraising structural interpretations using seismic data, with a strict focus on misfit functions. Misfit functions are a key component of struc- tural model appraisal. Understanding the relation between misfit functions and structural models is the first step towards building structural models that are optimally constrained by seismic data, and therefore less prone to interpretation uncertainties; it is also the first step towards automated interpretation quality control of structural models. We formalize the concept of appraising struc- tural models using seismic data. We start by giving a formal definition of what it means to appraise structural models using data; from this definition, we derive a set of conditions that are imposed on data space misfit functions for such an appraisal to be considered reliable. There are two major development in this paper. First, unlike in our previous studies, the conditions presented here no longer depend on some unknown reference model. Second, we show that there exist a special case of practical interest where all the conditions needed for appraising interpretations using vertical seismic profiling data can be satisfied }
}