To complete the new E16 from Kløfta to Kongsvinger, some 50 km NE of Oslo, 32 km of new motorway will connect Nybakk and Slomarka (Figure 1). In this project, NGI supplies the geotechnical design as a sub-contractor to COWI. NGI carried out an AEM survey using a Danish SkyTEM system with the aim to map depth to bedrock and get additional information about the extent of sensitive clay in the area.


A total of 178 line-km were flown in three consecutive days in January 2013. Three parallel lines with a spacing of 25 m were flown along the planned road corridor. In addition, 15 lines were flown near Vorma/Vormsund and 9 lines near Uåa, to cover potential areas of sensitive clay. These additional lines had a nominal spacing of 125 m.

Kart 1

Figure 1: Survey area 50 km NE of Oslo showing flight lines (red) and boreholes completed at the time of the survey (green). Thin white lines mark power lines in the area.
The survey area is characterised by a rather complex geology and highly variable bedrock depth. There is a trend towards shallower bedrock to the northeast whereas the areas around the rivers Vorma and Uåa are characterised by clay layers of up to 50 m thickness. In these areas possible occurrences of sensitive clay have been observed. 


AEM resistivity sections along the flight lines clearly trace the bedrock topography and in some areas indicate potential sensitive clay occurrences. Marine clays are characterised by resistivities between 1-10 Ωm (blue colours in Figure 2). The resistivity for sensitive clay is strongly site-dependent and can range from 10-100 Ωm (green colours in Figure 2). Other geological materials can have similar resistivity as sensitive clay, thus it is not possible to detect sensitive clay based on resistivity alone.


Figure 2: Typical resistivity model derived from AEM data integrated with borehole results. The red line depicts the manually picked bedrock layer. The black line indicates the 100 Ωm threshold. Blue colours in the boreholes indicate sedimentary material that is not sensitive clay, green colours indicate sensitive clay. Boreholes are marked by their numbers (4 digits) and their lateral distance to the AEM profile in meter.
To extract a 3D bedrock model from the AEM results existing bedrock tracking algorithms had to be further developed. As a first approximation, a predefined threshold resistivity (in this case 100 Ωm, based on visual inspection of borehole results) is tracked throughout the 3D resistivity  model. The depth to this layer is then assumed to represent the depth to the bedrock interface. Such an approach is usually successful for data within an area of homogeneous geology but the algorithm has only limited success for the extent of the entire survey.

For a survey of this extent over varying geology, the simple approximation of one threshold bedrock resistivity is not sufficient. NGI therefore developed an algorithm that first determines a spatial threshold resistivity model based on available borehole data and then applies this resistivity model to track bedrock between borehole locations. The final result is a bedrock model that agrees with boreholes and "fills in the gaps" where no borehole data is available.

Bedrock AEM

Figure 3: Depth to bedrock in the central survey area obtained from a tuned interpolation algorithm using AEM and borehole data. The grid resolution is 10 m.
Even though the AEM survey was carried out rather late in the project, it was possible to regain the survey costs through savings in the ground investigation programme as numerous planned boreholes could be omitted as the AEM bedrock model provided sufficient data. For future projects, we recommend to acquire airborne geophysical data early, in the ground investigation-planning phase, to both accelerate the investigation programme and to significantly reduce drilling COSTs.