Geostatistics and Clustering for Geochemical Data Analysis

• Author / Creator

  Prades, Carlos

• This thesis addresses challenges in geostatistical analyses of multivariate geochemical data that commonly contain complexities that have a significant influence on geostatistical modeling and cluster analysis. For geostatistical modeling, the effect of the most common despiking methods is investigated and their problems documented. It is shown that both local average despiking and random despiking lead to bias in the observed variogram and predicted uncertainty. A new despiking method is proposed and implemented to improve variography when the variable has a significant spike. The developed approach combines a random despiking component and a local average despiking component. Cluster analysis can be applied for mineral exploration purposes. It can be used to find large structures in the data and also to detect multivariate anomalous samples. Data transformations are shown to have a significant impact on clustering results. Guidance and recommendations on appropriate data transformations for improving cluster analysis performance are provided. Three different methods are developed for identifying multivariate anomalies with cluster and spatial analysis. The first method uses different combinations of clustering and data transformations for finding small anomalous clusters. The second uses different clustering outputs for identifying samples that do not clearly belong to any cluster. The third recognizes samples that are spatially anomalous. Each of these multivariate methods detects anomalies from a different point of view. A combination of these detection methods is recommended. The goal is to obtain more stable and reliable results. Its application in stream silt samples from the Northwest Territories shows that the proposed multivariate anomaly detection methods are capable of identifying several showings (known mineral deposits). Some of these showings are not detected from the histograms of different elements; this supports and motivates the use of multivariate anomaly detection methods for mineral deposit exploration.

• Subjects / Keywords

  • Geostatistics
  • Geochemical analysis
  • Clustering

• Graduation date

  Spring 2018

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/R3RN30N3B

• 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.