Inference of Locally Varying Anisotropy Fields

• Author / Creator

  Lillah, Maksuda

• Considering nonstationary features in geostatistical modeling is important. Recent developments in nonlinear estimation provide a practical way to infuse geological realism into numerical modeling by incorporating complex spatial features. Second order nonstationarity when present must be considered for more accurate geo-models. A critical step when considering second order non-stationarity is a parametric representation of the underlying anisotropy field quantifying the orientation and magnitude of anisotropy exhaustively. The inference of a locally varying anisotropy (LVA) field is deposit and data specific and this work presents a suit of methodologies for field generation in both 2D and 3D cases. Three different data types are explored: (1) 3D point data from direct angle measures from down bore hole formation etc, (2) exhaustive data from outcrop images, (3) compact geologic bodies typically inferred by geologists. Proposed methodologies for the generation of LVA field include reorienting point data by angle variance criteria and estimation by kriging, interpolating orientation in 3D by quaternion averaging, identifying local structures from exhaustive data by moving window technique and centerline extraction of geo bodies by thinning. Features of interest can occur at a coarser or finer scale than the grid size of the final model. Local structures of LVA can be better captured by allowing search areas to adjust to the range of the relevant features. In deposits characterized by varying scales of anisotropy, adaptive moving windows are used to assess the local orientation.
  
  Usually several different data types may need to be consolidated for the generation of a single LVA field and a tensor based combination scheme is discussed whereby multiple LVA fields can be merged. In the last section, data from a copper porphyry deposit is considered for LVA generation and results show higher correlation and lower mean squared error in cross validation analysis when kriging with LVA compared to traditional kriging.

• Subjects / Keywords

  • Geostatistics
  • Nonstationary
  • Anisotropy

• Graduation date

  Fall 2014

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/R3S10B

• 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.