Filling the gaps: Why geotech and geophysics are better together

Clouds roll over Austria’s highest peak revealing it part by part. Andi A. Pfaffhuber has made a career piecing together a more complete view of the earth by combining geotechnical and geophysical data.

Andi, CEO and co-founder of EMerald Geomodelling, is visiting the Austrian Alps. His team works on projects across Europe, helping clients minimise risk on civil engineering projects.

Historically, there have been misunderstandings between geotechnical engineering and geophysics teams. But, companies like his are helping both fields benefit by filling in each other’s data gaps.

Andi uniquely understands both perspectives: he completed a PhD in geophysics before diving into the engineering world at the Norwegian Geotechnical Institute.

“At NGI I was running a group that was developing and adapting geophysical methods, remote sensing methods, and GIS for geotechnical work.”

All experts were working together at the institute, but he discovered that it was a slightly different story in the field.

“There is a pretty big gap between geophysical and geological and geotechnical data. You usually have these different domain experts that work very independently,” says Andi.

“It’s unfortunately very common that the geophysicist doesn’t understand the engineer’s needs, and that the engineer lacks insight into what the geophysicist can do.”

Geotechnical data, like samples and boreholes, is key to confirming geological insights. Yet, collecting it can be expensive, time intensive, and only provides accurate information about that one location.

“You have some boreholes where you know exactly what’s going on in that borehole, but you know nothing about conditions in between them. Geophysics can fill those gaps,” says Andi.

Andi noticed geotechnical engineers were previously trying to fill in those gaps by drawing lines between boreholes. The process took a lot of time and created a static model, unless you started from scratch each time new data arrived.

These models made assumptions about the complexity of the surrounding geology, sometimes with costly consequences.

“If you assume your tunnel is in rock, but then it turns out that you actually don’t have rock cover because your assumed bedrock typography was not as high as you thought – that’s a big problem. And it’s unfortunately a problem that happens too often.”

The EMerald Geomodelling team uses airborne geophysical surveys that cover a wide area quickly and then only plan drilling projects where the geophysical data has high uncertainty.

“In most geological settings, the resistivity models that you get from airborne EM give you a good indication of where the bedrock interface is.”

Targeted boreholes can then validate the geophysical model or add new information. The end result is often a more accurate model in less time, and with less drilling costs.

Airborne electromagnetic (EM) surveys can quickly provide a good overview of geological conditions between boreholes.

But, like boreholes alone, the technique is not perfect. There’s one key aspect of geophysics data that’s been challenging for geotechnical engineers to work with: uncertainty.

“The concept of uncertainty is still difficult to handle. Geotechnical engineers traditionally prefer one line, although they know it won’t give the full picture,” says Andi.

He’d also heard stories from working alongside geotechnical engineers, where they were either oversold the certainty of geophysical models or the uncertainty in the data wasn’t clearly communicated.

“A super important factor in geophysical inversion is that you have a good way of treating the uncertainty in your measured data,” he says.

“There’s a number of people out there that we need to re-convince with new technology and an appropriate level of trust and openness about its limitations.”

The EMerald Geomodelling team creates visual ground models that include measures of uncertainty. By making uncertainty more tangible, geophysics can be easier for engineers to interpret and see the value in.

Recently, Andi’s team was brought into a highway project that already had a tunnel planned and was about to begin drilling. They asked EMerald Geomodelling for a second opinion to confirm their tunnel route.

“The bedrock topography could be gathered just a few weeks after we did the airborne TEM survey and the tunnel that was planned to be in rock turned out to be in soft clay,” says Andi.

“That was revealed without putting a first borehole in. So, the alignment of that tunnel could be changed right away.”

Before spending on drilling to discover the clay, geophysics offered a useful overview of the geology. After the tunnel was replanned, drilling was used to adjust and confirm the initial model.

“The project owners could see not just that they could save a lot of time and money in boreholes, but also that they could significantly decrease the geological risk in getting an overview much more quickly.”

It’s not the first time. His team helped one Norwegian highway project decrease drilling costs by 30% and time by up to six months (50%).

In the past, when a bedrock model was created – either geophysical or geotechnical – it was handed off as essentially complete. Using cutting-edge software and machine learning, the EMerald Geomodelling team has found ways to build more living models.

They use Workbench software’s specialised airborne EM tools to create geophysical models. Then, they combine those with borehole data using their own AI technology.

Their AI is trained to process, read, and calculate uncertainty in both the geophysical and geotechnical data. From there, the program generates a bedrock model that accounts for the gaps in both.

Because the AI is already programmed for the project, new data can be added and a new model generated almost instantly.

“We have an update functionality in our models. When the projects come back with a new set of boreholes, we just run the algorithm again. And the model is improved and gets more accurate.”

By working with a living model from day one, engineers and project managers can really appreciate the difference from traditional static models.

“We provide them with information very early, very quickly. But as the project goes along, we update that whenever they have new data, and then they’re super happy,” says Andi.

“Because they’re always working with a model that contains the best information that is out there.”

Andi’s key message isn’t that there’s one perfect method, it’s that geophysical and geotechnical data together build more accurate bedrock models.

“That’s the big value proposition for most geophysical methods: that you then can fill in the gaps between boreholes and get a more complete picture, a more complete overview, in a more efficient way.”

That’s how the EMerald Geomodelling team parts the clouds on the subsurface for clients, filling in one gap at a time.