Aberle, Michael1, de Caritat, Patrice2,1, McQueen, Ken1, Hoogewerff, Jurian1
1National Centre for Forensic Studies, Faculty of Science and Technology, University of Canberra, Canberra, Australia; 2Geoscience Australia, GPO Box 378, Canberra, Australia
Topsoil is a common material that may be transferred to people and objects prior to, during, or after perpetrating a criminal activity. Traditionally the use of soil material in law enforcement operations involves one-to-one comparison of recovered soil evidence with reference samples collected from known areas of interest. In casework and intelligence applications where this contextual information is not available, properties of the recovered evidence may be used to triage geographical regions as areas of low and high interest. If sufficient relevant spatial information is available, this approach may provide forensic intelligence to better focus operational resources on areas of interest. Here, spatial variability is both a strength and a weakness, with research required to determine 1) which parameters are sufficiently discriminatory at the chosen spatial scale, and 2) the effect of contributing anthropogenic and geogenic sources on forensic soil provenancing applications.
To investigate these issues, a high-density (1 site/km2) geochemical survey has been conducted to map the compositional variation of ~700 urban topsoil (0-5 cm depth) samples across Canberra, Australian Capital Territory. As a study location, Canberra represents a large city (population > 450k) with a system of urban open green spaces and bushland reserves, as well as minimal heavy industrial activity. Thus, in addition to law enforcement applications, the effect of urbanisation on the environment can be studied without significant masking by industrial pollutant sources. Using standard protocols for urban geochemical surveys, including extensive quality control measures, the samples have been prepared in a fusion flux matrix and characterised for bulk elemental geochemistry using X-Ray Fluorescence and Inductively Coupled Plasma – Mass Spectrometry analysis of total acid digests. Bulk mineralogy has been determined on a subset of samples using powder X-Ray Diffraction.
The results demonstrate that the geochemistry of the topsoils is strongly influenced by the dominant lithological units of the underlying bedrock. While the concentrations of known anthropogenic elements are mainly below thresholds for health-based investigation (with 1 or 2 sites marginally at threshold), there is evidence of diffuse anthropogenic contributions, particularly in older suburbs and light industrial areas.
For provenancing applications, a number of different approaches have been suggested to ‘match’ a recovered sample to a map. Typically, these involve comparing the properties of the recovered sample to those in each ‘target’ survey grid cell and attributing statistical significance to some measure of ‘overlap’ at an arbitrary inclusion/exclusion threshold (e.g. 95%). By instead comparing each grid cell to a series of questioned samples from within and outside the survey boundary, and weighing each probability by the probability of observing the grid cell value in the total dataset in a likelihood ratio approach, we have demonstrated that regions of interest may be reduced in a more conservative manner better suited to forensic provenancing than previous approaches.
Further merits and challenges of provenancing topsoil from urban environments will be presented, notably the significance and impact of displacement and introduction of topsoil material from other areas on developing forensic soil surveys, as well as determining the source of questioned samples.
Michael Aberle is currently a PhD candidate at the National Centre for Forensic Studies, University of Canberra. Together with industry partners, his research primarily focuses on evaluating the forensic utility of a fit-for-purpose, high-density soil geochemical survey of the Australian Capital Territory for forensic provenancing of bulk and trace soils.