Combining Three Different Methods of Soil Characterization.

Soil samples are a common trace in forensic casework, but they are not always used as evidence. This is partly due to the lower evidential value of soil as compared to for example DNA and partly because soil is less well known for having a forensic use. So, in order to use soil samples more frequently the evidential value has to be increased. Soil samples can be used to determine whether or not two samples are from the same source. By doing so a piece of evidence can be linked to a location. Soil comparisons can be performed using different components in the soil where each component has its own resolution, advantages and disadvantages. By combining measurements of different independant components in the soil samples the comparison becomes more robust and the evidential value of the comparison can increase. Here we report the use of three different methods: The bacterial DNA profile of the soil samples, which is highly discriminative, uses the bacteria present and needs only a small amount of sample material. Elemental composition, which has a lower resolution, uses the mineralogical composition and needs a larger amount of sample material, but is non-destructive. Pollen, which is very discriminative, uses the pollen grains present which reflect the current and the old vegetation and needs only a small amount of sample material. For all three methods we quantify the similarity between samples using a statistical distance. We then use a database to quantify how rare this similarity is for soil samples from the same location and soil samples from different locations. This approach equals the Bayesian approach, in that it gives a measure of the likelihood of finding a certain similarity given that either the samples came from the same source or the samples came from different sources. This makes it possible to combine the results of the different methods objectively. The combining of methods is demonstrated in two case examples.