Soil is an inherently heterogenous medium and characterization of site soils is subject to various scales of variability. Often practitioners automatically turn to table E1 in AS4482.1-2005 for guidance on how many samples they need to identify ‘hotspots’ at a site. This may not be appropriate as soil contamination may not be present in discrete areas and almost never as a circular shape. Practitioners should consider more robust statistical methods for determining an appropriate sample quantity and also analyzing their data.
What was the "light bulb" or most radical/innovative point?
Often an analysis of a very small quantity of soil ends up being used to characterize large volumes of soil. The chances of a 10 g final analytical sub-sample being representative of the mean field sample concentration and this in turn being representive of the mean contaminant exposure concentration for potentially hundreds of cubic metres of site soil would be the equivalent of winning the lottery. A robust understanding of short-scale spatial heterogeneity and micro-scale (intra-sample) heterogeneity will ultimately aid site characterization and potentially lead to large savings to the client.
Was anything controversial or contentious and what were the different perspectives?
It was suggested that, if used blindly, the application of AS4482.1-2005 can cause more trouble than it’s worth. Whilst guidance can be useful people can get lazy and not apply the level of scientific rigour as to whether the basis for the guidance they are applying is appropriate for the problem at hand.
What was unresolved?
The Australian guidance for statistically robust site characterization is limited. Methods such as the Incremental Samplng Methodology introduced in the USA by the ITRC may provide a way forward but regulatory understanding/approval may limit its current application.
How can things be done better as a result of this discussion?
Practitioners can take simple steps for major gain. Basic application of the PSI process and the fundamental statistics for sampling design and interpretation will go a long way to curing many issues observed in the industry.
Didn't get to attend the event, but would love to review the presentations - go to the ALGA online library* type in search words "consultants | court"
21 March 2019 event report by Jeremy Richardson
* note the access to all the papers in our online library is a member only benefit, for more information or to join click here