What can we learn from assisted stratigraphic correlation ? A case study from the North Sea

Jonathan Edwards and Victor Haddad and Guillaume Caumon and Cedric Carpentier and Silvan Hoth and Marcus Apel. ( 2017 )
in: 2017 Ring Meeting, pages 1--14, ASGA

Abstract

Improvements have recently been made to manage uncertainty during automatic stratigraphic well correlation. While the algorithm is automatic, it is still up to the geologist to interpret the raw data to identify stratigraphic units, to choose the correlation rules, and to critically analyze and interpret the correlation results. In this paper, we confront different stratigraphic correlations ob- tained by different methods on well logs interpreted as sets of transgressive and regressive sequences (with and without biostratigraphic constraints, and with and without one-to-many matching). We observe that the one-to-many matching mitigates the impact of transgression/regression sequence identification uncertainties, and that biostratigraphy is a key information both for assessing the reliability of the results and for constraining the model space.

Download / Links

BibTeX Reference

@INPROCEEDINGS{Edwards2017,
    author = { Edwards, Jonathan and Haddad, Victor and Caumon, Guillaume and Carpentier, Cedric and Hoth, Silvan and Apel, Marcus },
     title = { What can we learn from assisted stratigraphic correlation ? A case study from the North Sea },
 booktitle = { 2017 Ring Meeting },
    number = { 1949 },
      year = { 2017 },
     pages = { 1--14 },
 publisher = { ASGA },
  abstract = { Improvements have recently been made to manage uncertainty during automatic stratigraphic well correlation. While the algorithm is automatic, it is still up to the geologist to interpret the raw data to identify stratigraphic units, to choose the correlation rules, and to critically analyze and interpret the correlation results. In this paper, we confront different stratigraphic correlations ob- tained by different methods on well logs interpreted as sets of transgressive and regressive sequences (with and without biostratigraphic constraints, and with and without one-to-many matching). We observe that the one-to-many matching mitigates the impact of transgression/regression sequence identification uncertainties, and that biostratigraphy is a key information both for assessing the reliability of the results and for constraining the model space. }
}