Date of Graduation
5-2025
Document Type
Thesis
Degree Name
Master of Science in Geology (MS)
Degree Level
Graduate
Department
Geosciences
Advisor/Mentor
Sharman, Glenn R.
Committee Member
Dumond, Gregory
Second Committee Member
Shaw, John B.
Keywords
Geology; Mineralogy; Multidimensional Scaling; Principal Component Analysis
Abstract
Methods for determining and categorizing the modal compositions of sand and sandstone have long been a subject of debate in the field of sedimentary geology. Point counting is the most commonly used technique for determining modal compositions from petrographic slides, which are then categorized based on relative proportions of quartz, feldspar, and rock (or lithic) fragments. However, this approach fails to adequately preserve relevant textural data, such as grain size and sorting, which play an important role in diagenesis and in influencing reservoir quality. Additionally, the categorical nature of point counting results in a lack of specificity and loss of valuable data regarding intra-grain mineral proportions, which could be important for provenance analysis.
The Tescan Integrated Mineral Analyzer (TIMA) can address some of the deficiencies associated with traditional point counting by providing discrete mineral maps of individual grains within a sample. TIMA is a fully automated analysis, during which a scanning microscope utilizes backscattered electron (BDE) and energy dispersive X-ray (EDX) technology to determine the precise elemental makeup of a sample. This information may then be used to infer mineralogy at micron resolutions.
This study explored the insights gained from TIMA data for determining the grain size and modal composition of seven modern sand samples with a diverse suite of compositions and textures. The purpose of this work was to develop a quantitative approach to analyzing the composition and morphology of clastic sediment to provide more insight than point counting alone in hopes of providing value to future provenance studies. A combination of ImageJ and Python scripts was used to analyze the mineralogic composition and morphology of individual grains and the distribution and compositions of each grain size within a given sample. I applied nonmetric multidimensional scaling using two dissimilarity metrics to evaluate inter-grain mineral relationships within a sample. I then assessed the quality of the fit of these solutions and suggested several explanations for suboptimal model performance. Additionally, I performed principal component analysis as an alternative method to evaluate inter-grain mineral relationships and evaluated the effectiveness of that approach. I found that, though conventional metrics suggest a poor fit, meaningful relationships could be drawn from multidimensional scaling analysis. I also concluded that principal component analysis may be a better choice for visualizing and interpreting intra-grain mineral relationships. Finally, I found that morphological analysis is both possible and insightful with TIMA data. Ultimately, this work serves as a notable step toward leveraging TIMA data for provenance analysis of sand(stone) composition.
Citation
Jones, Q. (2025). Insights From Automated Mineralogic Analysis of Modern Sand. Graduate Theses and Dissertations Retrieved from https://scholarworks.uark.edu/etd/5724
