Image of Estimating uncertainties in 3-D models of complex fold-and-thrust belts: A case study of the Eastern Alps triangle zone

Text

Estimating uncertainties in 3-D models of complex fold-and-thrust belts: A case study of the Eastern Alps triangle zone

Geological modeling commonly results in a single prescribed geometric representation of the subsurface with no consideration of uncertainties. Accounting for uncertainties is of particular importance in the triangle zone at the leading edge of deformation of the foreland fold-thrust belt of the European Alps, the Subalpine Molasse. Here, interpretations of the complex structures are limited to 2D and are based almost exclusively on subsurface data, lacking the constraint of surface data.
Implicit modeling can be used to create automated 3D model realizations considering parameter uncertainty. In this sense, multiple possible models can be assessed within the uncertainty range assigned. As implicit modeling often yields artifacts or geologically unfeasible scenarios, the concept of geological topology can be used to constrain the modeling outputs, so that only models without artifacts are considered into the final model ensemble.
Two experiments are designed to test out this workflow. The methodology is first tested using a very simple synthetic model, and later applied to a portion of the Subalpine Molasse, where different 2-D subsurface geometries have been proposed. For the first time, a 3D implicit model ensemble incorporating a full assessment of uncertainties is built in this area. Results show that it is feasible to use topological information for posterior model constraint, even in models of high structural complexity. To eliminate all non-meaningful models, however, it is necessary to use more than one topological constraint. A comparison of the prior and posterior model ensembles shows a correlation between model parameters and a shift in parameter probability density curves in the synthetic model experiment, and a decrease in entropy for both experiments. A topological constraint should thus be applied routinely when building a stochastic implicit model ensemble.

Availability
147551.136Perpustakaan BIG (Eksternal Harddisk)Available
Detail Information
Series Title
Applied Computing and Geoscience - Open Access
Call Number
551.136
Publisher
Amsterdam : Elsevier.,
Collation
16 hlm PDF, 11.035 KB
Language
Inggris
ISBN/ISSN
2590-1974
Classification
551.136
Content Type
text
Media Type
-
Carrier Type
-
Edition
Vol.18 June 2023
Subject(s)
3D implicit modeling
Uncertainty topology
Eastern Alps
Triangle zone
Specific Detail Info
-
Statement of Responsibility
Other version/related

No other version available

File Attachment
  • Estimating uncertainties in 3-D models of complex fold-and-thrust belts: A case study of the Eastern Alps triangle zone
    Geological modeling commonly results in a single prescribed geometric representation of the subsurface with no consideration of uncertainties. Accounting for uncertainties is of particular importance in the triangle zone at the leading edge of deformation of the foreland fold-thrust belt of the European Alps, the Subalpine Molasse. Here, interpretations of the complex structures are limited to 2D and are based almost exclusively on subsurface data, lacking the constraint of surface data. Implicit modeling can be used to create automated 3D model realizations considering parameter uncertainty. In this sense, multiple possible models can be assessed within the uncertainty range assigned. As implicit modeling often yields artifacts or geologically unfeasible scenarios, the concept of geological topology can be used to constrain the modeling outputs, so that only models without artifacts are considered into the final model ensemble. Two experiments are designed to test out this workflow. The methodology is first tested using a very simple synthetic model, and later applied to a portion of the Subalpine Molasse, where different 2-D subsurface geometries have been proposed. For the first time, a 3D implicit model ensemble incorporating a full assessment of uncertainties is built in this area. Results show that it is feasible to use topological information for posterior model constraint, even in models of high structural complexity. To eliminate all non-meaningful models, however, it is necessary to use more than one topological constraint. A comparison of the prior and posterior model ensembles shows a correlation between model parameters and a shift in parameter probability density curves in the synthetic model experiment, and a decrease in entropy for both experiments. A topological constraint should thus be applied routinely when building a stochastic implicit model ensemble.
Comments
You must be logged in to post a comment