A Methodology to "Unmix" Bulk Elemental Concentration Data of Mudstones Using Positive Matrix Factorization
Abundant XRF data for mudstones is currently available due to recent innovations in analytical technology and techniques for hydrocarbon extraction. However, advanced statistical analysis of these datasets could reveal stratigraphic trends previously hidden to traditional techniques. This study developed a method of statistically grouping the concentration and stratigraphic distribution of elements within geological formations using positive matrix factorization (PMF) of portable x-ray fluorescence (pXRF) data from drill cores. Environmental Protection Agency: Positive Matrix Factorization (EPA PMF) is a multivariate receptor model typically used for quantifying the contribution of particulate matter aerosol sources to samples based on the composition or “fingerprints” of the sample. A workflow was developed to utilize EPA PMF for a geologic application using pXRF data from mudstones. The PMF workflow was applied to data obtained from pXRF analysis of three drill cores from New York State containing Ordovician carbonates and mudstones. Results from PMF were compared to previous interpretations made using traditional sedimentology methods. PMF produced lateral and stratigraphic variability, lithologies, sediment provenance, and paleo-redox conditions that matched previous work and revealed new trends unseen by earlier studies. PMF proved to be an effective tool for quantitatively analyzing elemental proxy data based on statistical analysis and identifying stratigraphic variability. The PMF workflow produced more objective and defendable interpretations, added credence to existing work, and highlighted previously unidentified stratigraphic trends. XRD analysis is used to address mineralogy complications due to pXRF detection limitations and SEM analysis will distinguish between detrital and biogenic silica.
AAPG Datapages/Search and Discovery Article #90351 © 2019 AAPG Foundation 2019 Grants-in-Aid Projects