Comparison of 2D and 3D uncertainty quantification in geostatistical modeling in stratigraphic formations

• Author / Creator

  Latifi, Amir Mahdi

• Under suitable geological conditions and depending on the modeling goals at the project’s current phase, a geological modeler could justifiably choose to estimate natural resource concentrations by building geostatistical models on a 2D grid rather than a 3D grid. While the application of these 2D models is limited only to resource estimation, they take less time for a modeler to build and while generating similar overall results to 3D models, are simpler to build. Thus, in geological domains that have a flat and layer-like geometry, i.e., tabular mineral deposits, modelers might prefer these 2D models to 3D models.
  While the implementation of 2D geostatistical modeling in these geological domains is well established, the quantification of uncertainty in these workflows is not, especially in comparison to 3D models in the same domains. Since the uncertainty in final resource estimations is a critical economic factor in deciding the feasibility and risk of investing in a mineral/hydrocarbon extraction campaign, an understanding of differences in uncertainty quantification between 2D and 3D models allows modelers to make informed choices on which type of modeling workflow is most appropriate for their specific project.
  To make a comprehensive comparison of uncertainty quantification in 2D and 3D modeling, all aspects of uncertainty are considered and grouped into three categories: uncertainty in model parameters, uncertainty in the geometry, and residual uncertainty (which is the uncertainty in the models that include no parameter uncertainty). These categories are analyzed through two lenses: firstly, by building 2D and 3D models in the McMurray formation of Northern Alberta as a case study and analyzing the results, and secondly, by looking at each category separately through analytical tests.
  The results show that the smaller size of the 2D domain and data set leads to higher uncertainties in some categories, such as histogram uncertainty or residual uncertainty. However, this is not necessarily true in other categories, such as the uncertainty in the geometry, and uncertainty could be higher in the 2D or 3D workflow depending on modeler choices and domain characteristics. In the case study, the 3D workflow shows almost identical overall global uncertainty in the results compared to the 2D workflow. This emphasizes the importance of incorporating full uncertainty quantification in the results to have realistic resource estimations while indicating the loss of information in the 2D workflow that causes higher uncertainties in some aspects of the workflow. Additionally, the results from analytical tests provide some insight into contributing factors to differences between uncertainty quantification in 2D and 3D workflows, such as the vertical size of the domain, degree of vertical spatial continuity, and sample density that have an impact on the degree of differences between the two workflows. By considering these factors, the modeler could make a more informed decision on the type of workflow for their project with an understanding of probable differences between the two workflows.

• Subjects / Keywords

  • Geostatistics
  • Mining
  • Resource modeling
  • Uncertainty quantification
  • Resource estimation

• Graduation date

  Fall 2023

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/r3-8n6p-4x34

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.