"Smart sampling" trial over the Nullarbor's "blue hole"
The famous "blue hole" – also known as the Coompana magnetic anomaly – on the South Australian section of the Nullarbor Plain, gets its name from the colour generated by scans of what is an area of deep and unusually low magnetism. It has fascinated geologists since first noticed more than 40 years ago.
A recent "smart sampling" trial undertaken over the blue hole was not specifically designed to test it, though the surface geochemical and vegetation sampling might help throw more light on the geological mystery.
The primary aim was to trial a system of rapid, in-field sampling and analysis using helicopter transported scientific facilities to characterise the chemistry of a large area. The work may lead to more detailed exploration planning.
Any additional knowledge of the blue hole would be a bonus and perhaps help fine-tune theories about its genesis, which include being formed when the Earth's magnetic field was reversed, following a meteor impact, or because a northern hemisphere remnant shifted south by tectonic movement.
Collaborating with the Geological Survey of South Australia
The team comprised of researchers from the Geological Survey of South Australia (GSSA) and CSIRO, with assistance from two traditional land owners and a helicopter crew. The field trip was designed to outline the chemical characteristics of a 4000 square kilometre (km) area, an expanse roughly twice the size of Melbourne and its suburbs – in a week.
Traditional soil sampling can take months using ground-based crews, as individual samples weighing two to four kilograms are collected and returned to a central laboratory for analysis. This trial replaced traditional techniques with advanced portable tools, such as x-ray fluorescence (XRF) and spectral analysis.
Samples of vegetation, mainly bladder salt bush and pearly blue bush, were also collected and tested for the presence of minerals.
In several senses it was a project "on the fly" as they travelled via helicopter to 310 sites that they subjected to surface geochemical sampling and analysis.
In-field analysis and machine learning
Ryan Noble, a principal research scientist with CSIRO and member of the team, says the trial was remarkably successful.
"The starting point was to take sufficient soil and rock samples to characterise the area, which meant digging a small hole between 20 and 30 centimetres deep at each sample point, as well as collecting vegetation," Dr Noble says.
"Samples were prepared in the field, which was one of the more unique features of the project, because it involved crushing and milling and then pressing each sample into a form suitable for portable XRF and spectral analysis."
XRF provided a good indication of chemistry, spectral analysis provided a proxy for the mineralogy.
Rapidly surveying a large geographic area
The entire process was rapid with an analysis achieved every four minutes thanks to the creation of an efficient production line, an essential aspect of a project visiting so many sites in an area measuring 80km by 50km.
As well as the broadacre survey based on 4km centres, the focus was narrowed down to take a close look at a specific area with samples of 36 centres in a hexagon shape with 1km centres.
The smaller area was chosen partly because of interesting results, but also to test how it could be done by mineral explorers working in an unknown area with the advantage of having in-field tools. This would enable a specific area to be analysed quickly and in greater detail without having to organise a return field trip, which is what mineral explorers currently have to do.
Alongside its partners, CSIRO is showing how new technology and research solutions can be tailored to give organisations the cutting edge necessary to grow and compete in an ever more competitive environment.
"What we showed is that we could conduct a regional sampling operation, generate good geochemical and mineralogical results, and then refine our sampling while on site by returning on the last day to an area of particular interest," Dr Noble says.
The first lesson learned from taking the latest in science to the outback, rather than simply retrieving samples of the outback for later scientific analysis, was that it can be done, and a lot quicker than was expected.
The second was that the silica content in the soil samples appeared to match up with the blue hole magnetic anomaly, but was probably related to a limestone unit closer to the surface.
"We learnt a bit about the area that we didn't know before. We learnt how to survey much more quickly and that could represent a big cost saving when applied by a mineral exploration company," Dr Noble says.
"We have more to learn. One improvement that might become clearer after we go back over the modelling is to see whether we could have got away with only collecting 80 samples instead of the 300, which is very much what smart sampling is about.
Plans to refine the method
"The Coompana field trip was a test case to demonstrate that an exploration project can cover an area much more efficiently with new technology over traditional methods that can involve generating thousands of samples and take months, not days."
Refinements are planned to the smart sampling method, such as how much material is needed in each sample, to replace the two-to-four kilograms collected in a traditional survey with as little as 200 grams, a major weight and cost saving when a helicopter is used.
"Now that we've characterised the area, we have to take the next step which is looking for an anomaly which could become an exploration target. We hope to do that with the GSSA in an area with known mineral deposits and expand from there," Dr Noble says.
"As we build the knowledge base by testing an area with known mineralisation we can apply our model to that data, and use that to find other mineralised areas."