Putting the geo into geophysics

I have just taken a job. You might have heard of the company, Fleet Space Technologies. They're based in Adelaide, Australia, but they have a global team. And they are fast becoming THE experts in a particular geophysical methodology - Ambient Noise Technology, ANT.

The first time I heard about ANT was when a remarkable seismologist, Dr JP O'Donnell was working at the Geological Survey of South Australia, as was I. He explained the method utilises "noise", that is the background hum of the Earth and some sophisticated mathematical deconvolution programming to render a signal out of that noise. Things like the movement of waves, or wind, or cars or trucks and trains, basically anything that interacts with the ground creates noise and this can be used for imaging the Earth.

Rayleigh and Love wave empirical Green’s functions (EGF) for station pair SOC07-SOC13 (see Fig 1 for locations) determined by stacking 30 days of vertical and transverse component symmetric noise cross-correlograms, respectively. From O’Donnell et al 2020, MESA Journal.

Now, JP was mostly interested in the deep Earth, tens of kilometres, and so his work relied heavily on the reverberations of distant earthquakes as the noise from which he would resolve seismic signal to create images of the subsurface. These images are super important and interesting. They show that large swaths of the crust have unique seismic velocities, and structural features within the crust can be resolved.

It turns out that each signal can be cross correlated between the each geophone. The time taken for a given each noise "event" to travel between each geophone makes for the signal that tells the seismologist the velocity structure of the subsurface. A high frequency noise will be bouncing around in the shallow crust, and the speed with which it travels tells something of the density and temperature of the subsurface.

This is great for imaging large features with long wavelength signals.

But what about the other end of the spectrum? Can high frequency signals inform us about the shallow crust? Turns out the answer is yes, a resounding yes. And it is this near surface work that Fleet Space specialises in.

They use specially designed geophones to record signal over a survey area for a few days then produce a sophisticated 3D model of the velocity structure of the upper 2 to 3 km of the earth. In mining terms that is a deep look into the subsurface. Compared to the entire crust or lithosphere it's pretty shallow. I had a fun interaction with some of my new Fleet Space colleagues when they started to tell me how "deep" the ANT images. I laughed because to me a few kilometres is not deep. That's not deep at all when you consider the radius of the earth is nearly 6000km. But I know what they mean. And the 3D models they produce are really interesting!!

Slice through ANT velocity model from Hillside Cu-Au deposits . From Joneset al, 2024

Over the past year or so Fleet Space has been busy hiring geologists to build capacity to interpret their geophysical models. You see I've learned over time that doing geophysical surveys and producing geophysical models is one skill set. But interpreting those models in terms of geology, tats another. Sometimes people can do both. More often we each have our speciality. I tell you now, if the chat on the internal Slack channel at Fleet Space is anything to go by, I will never be able to build 3D geophysical models! The maths, the programming, the language is foreign. But, I can certainly have a go at interpreting those models.

So that's what I'm doing. Trying to work with the Fleet Space geology and geophysics teams to put the geo into geophysics and help the various clients of Fleet Space to better understand the meaning of the models they receive.

Like any geophysical method, ANT is sensitive to a particular physical property. In this case, seismic velocity - how fast sound travels through rock. So we always need to keep that in mind when looking at the models. We may not be looking at any particular structure, we might be just looking at an average velocity of a block of rock. The models need to be interpreted in the context of other geological and geophysical data sets. And therein lies the interesting part! I'm enjoying the work immensely so I'll get back to you sometime when I've written a paper or two. There's a couple baking now and I'm keen to share more as I learn more about the ANT technique and about the plans of Fleet Space as well. Apparently it's Australia's fastest growing company right now. Exciting times.