Part of Canberra, Australia, where scientists have been using geochemical analysis to develop a predictive soil provenancing method.
Scientists in Canberra, Australia, have been using geochemical analysis to develop a predictive soil provenancing method. Credit: Michael Aberle

In the very first appearance of Sherlock Holmes, 1887’s A Study in Scarlet, Dr. Watson jots down notes on the famous detective’s incredible powers of observation. Holmes “tells at a glance different soils from each other. After walks, has shown me splashes upon his trousers and told me by their colour and consistence in what part of London he had received them.” Holmes finds footprints in a claylike soil in the story and uses his knowledge of geology in several subsequent mysteries.

Today, scientists are trying to expand the envelope of forensic geology, the science of using unique characteristics of geological materials in criminal investigations, through more refined techniques.

Better Tools for Soil Sleuths

Law enforcement agencies often try to piece together the tracks of criminals by analyzing soil and dust samples left on items such as clothing and vehicles. The concept has been around at least since the time of Arthur Conan Doyle, who popularized it.

In a study published in the Journal of Forensic Sciences, however, researchers in Australia outline their efforts to develop a more powerful tool for detectives. In a nutshell, they took soil samples and subjected them to advanced analytical methods and concluded that the “empirical soil provenancing approach can play an important role in forensic and intelligence applications.”

The researchers looked at 268 previously collected soil samples, each from its own square kilometer in an area of North Canberra measuring some 260 square kilometers, more than 4 times the size of Manhattan Island. Geochemical survey data were mapped to show what the chemical and physical properties should be between measured points, as well as the measured uncertainty for each cell. Comparative samples were analyzed using Fourier transform infrared spectroscopy, magnetic susceptibility, X-ray fluorescence, and inductively coupled plasma–mass spectrometry.

The researchers were given three samples from the surveyed area and challenged to identify the 250- × 250-meter cells they came from. Using the method, they were able to eliminate 60% of the area under consideration. In a real investigation, that would mean spending less of law enforcement’s time and money on areas that won’t yield any results.

“This is extremely useful in forensic investigations as it allows [investigators] to prioritize resources to the most promising parts of the area.”

“We can use fairly standard analytical methods and achieve degrees of exclusion generally of the order of 60% to 90% of the investigated area,” said study lead author Patrice de Caritat, principal research scientist at Geoscience Australia. “This is extremely useful in forensic investigations as it allows [investigators] to prioritize resources to the most promising parts of the area. The greatest challenge was, predictably, the natural heterogeneity of soils. Even when characterized empirically at a density of one sample per 1 square kilometer, it is always possible that the evidentiary sample is uncharacteristic of that 1 square kilometer.”

De Caritat said predictive provenancing using existing digital soil maps has never been put forward before in a forensic application and offers an “effective desktop method of ruling out vast areas of territory for forensic search” as soon as a soil analysis is available. The research builds on an earlier paper he coauthored in which the method reduced the search area for soil samples by up to 90%.

“Using soil in forensic cases is an old technique,” added de Caritat. “Originally, bulk properties such as color were used and really useful, particularly to exclude matches between soil samples. But what has changed now is the breadth and depth of techniques brought to bear on the question.”

Hunt for Suspects

Lorna Dawson, head of the Soil Forensics Group at The James Hutton Institute in Aberdeen, Scotland, said the research was well carried out and is a good proof of concept but unrealistic because sampling at such a high resolution is not affordable.

“Often, the sample recovered from the car, tool, shoe, etc., is too small to allow elemental profile analysis to be carried out, so the methods described in Patrice’s research would not work in every country, and to test it would be prohibitively expensive,” said Dawson, who was not involved in the study.

“But if funding could be made available by some international donor, we as practitioners could work with key researchers such as Patrice, and we would be delighted to carry out the research to set up the appropriate databases and models to link with currently available soil databases,” Dawson added. “That level of detail would certainly help in many serious crime investigations such as fakes, frauds, precious metals, etc.”

De Caritat and colleagues have received a Defence Innovation Partnership grant from the South Australia government to apply their provenancing work to soil-derived dust and to include X-ray diffraction mineralogical and soil genomic information to increase specificity. The project is a collaboration involving the University of Canberra, the University of Adelaide, Flinders University, the Australian Federal Police, and Geoscience Australia. The research may be used for counterterrorism, where “environmental DNA” on the clothing or other personal effects of a suspect may prove valuable.

Jennifer Young, a lecturer in the College of Science and Engineering at Flinders University not involved in the new research, said last year that the technology could help provide “evidence of where a person of interest might have traveled based on the environmental DNA signature from dust on their belongings.”

—Tim Hornyak (@robotopia), Science Writer


Hornyak, T. (2021), Predictive forensics helps determine where soil samples came from, Eos, 102, Published on 16 August 2021.

Text © 2021. The authors. CC BY-NC-ND 3.0
Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.