Data interpretation in forensic sediment and soil geochemistry

Automated geochemical techniques enable reproducible elemental assays of small quantity samples and have been used in recent years in many forensic criminal investigations in England and Wales. Two case studies are presented which highlight the problems of testing the presence of pre-, synor post-crime event sample mixing. The number of elements or compounds analysed can have a bearing upon statistical discriminant techniques which may provide false-positive or false-negative associations or exclusions. Chemical analyses of soils and sediments using both Atomic Absorption Spectrometry (AAS) and Dionex, and Inductively Coupled Plasma Mass Spectrometry (ICP-MS) enabled the identification and classification of discrete groups by hierarchical cluster analysis and Canonical Discriminant Function Analysis. These groupings however, prove fragile to small variations within samples of even the most common minerals.