A probabilistic approach towards source level inquiries for forensic soil examination based on mineral counts

Area of Science:

• Forensic Science
• Geology
• Mineralogy

Background:

• Forensic soil examination is crucial due to soil's complex and varied nature.
• Mineral suite studies are common in soil analysis but lack statistical interpretation.
• Probabilistic approaches are needed to enhance the interpretative power of forensic soil examination.

Purpose of the Study:

• To develop and validate statistical methods for the probabilistic interpretation of forensic soil analysis.
• To fill the gap in statistical approaches for mineral suite studies in forensic science.
• To enable answering source-level questions in forensic investigations using soil mineralogy.

Main Methods:

• Microscopical analysis of light and heavy mineral fractions (90-180 µm) from Lausanne, Switzerland soil samples.
• Identification and counting of light minerals using Scanning Electron Microscopy with Energy-Dispersive X-ray Spectroscopy (SEM-EDS).
• Manual identification and counting of heavy minerals using Polarized Light Microscopy (PLM), followed by multivariate statistical analyses (PCA, HCA, LDA) and blind testing.

Main Results:

• Multivariate statistical treatments successfully identified pertinent variables for soil classification.
• Classification models were built and validated using blind tests, demonstrating their effectiveness.
• The study successfully demonstrated a probabilistic approach to interpreting mineralogical data for forensic soil examination.

Conclusions:

• Statistical and multivariate methods provide a robust framework for the probabilistic interpretation of forensic soil mineralogy.
• The developed approach enhances the ability to answer source-level questions in forensic investigations.
• This study establishes a foundation for more rigorous statistical analysis in forensic soil science.