Building stratigraphic models from sparse data

Jonathan Edwards. ( 2017 )
University of Lorraine

Abstract

All stratigraphic models building and analysis are based on stratigraphic correlations of sedimentary units observed on wells or outcrops. However, the geologist building these stratigraphic correlations faces two main problems. First, the data available are few and sparse. Second, the sedimentary processes leading to the deposition of the units are numerous, interdependent and poorly known. So, the construction of a stratigraphic correlation model might be seen as an under-constrained problem with several possible solutions. The aim of this thesis is to create a numeric method to generate stochastic stratigraphic models that are locally constrained by observation data. Two steps are necessary : 1. The establishment of rules describing the spatial organization of sedimentary units observed on outcrops and wells. For these rules, two axis are explored : | The formulation in equations of rules defined in the sequence stratigraphy framework. These rules, presented qualitatively in the literature are translated in quantitative terms to evaluate the probability of two sedimentary units to be correlated. | The deduction of the probability of two sedimentary units to be correlated from stratigraphic models built from forward stratigraphic methods. 2. The development of an algorithm to build possible stochastic stratigraphic models from the rules cited above and observation data.

Download / Links

BibTeX Reference

@PHDTHESIS{,
    author = { Edwards, Jonathan },
     title = { Building stratigraphic models from sparse data },
      year = { 2017 },
    school = { University of Lorraine },
  abstract = { All stratigraphic models building and analysis are based on stratigraphic correlations of sedimentary
units observed on wells or outcrops. However, the geologist building these stratigraphic correlations faces two main
problems. First, the data available are few and sparse. Second, the sedimentary processes leading to the deposition of
the units are numerous, interdependent and poorly known. So, the construction of a stratigraphic correlation model
might be seen as an under-constrained problem with several possible solutions.
The aim of this thesis is to create a numeric method to generate stochastic stratigraphic models that are locally
constrained by observation data. Two steps are necessary :
1. The establishment of rules describing the spatial organization of sedimentary units observed on outcrops and
wells. For these rules, two axis are explored :
| The formulation in equations of rules defined in the sequence stratigraphy framework. These rules, presented
qualitatively in the literature are translated in quantitative terms to evaluate the probability of two
sedimentary units to be correlated.
| The deduction of the probability of two sedimentary units to be correlated from stratigraphic models built
from forward stratigraphic methods.
2. The development of an algorithm to build possible stochastic stratigraphic models from the rules cited above
and observation data. }
}