Figure 1: Seismicity in The Netherlands between 1900 and 1996 (red circles). The Roer Valley Graben, in the South of The Netherlands, is bounded by the Peel Boundary Fault in the North and the Feldbiss Fault in the South. Both faults show a NW-SE trend and are denoted in black. The survey area, at the Peel Boundary Fault, is marked by the yellow square.

It follows from figure 1 that natural seismicity is mainly restricted to the Southern part of the Netherlands and relates to the Roer Valley Graben. This graben is bounded by the Feldbiss fault to the south and the Peel Boundary fault to the north. The Peel Boundary fault is still active and it was expected to be active in earlier, Quaternary, times. Present day activity is clearly showed by the Roermond earthquake of 13-4-1992, Ml = 5.8. Since the Feldbiss fault was already subject of paleo-seismological investigations (Camelbeeck and Megraoui, 1996 and 1998), the present study concentrates on the Peel Boundary fault.

• A ground water level lower than 4 meters to avoid expensive pumping during the opening of the trench.
• Being in an area where accurately dated Maas river terasses could be used for dating purposes in the trench.
• Having a laterally continuous fault to eventually be able to couple vertical displacement to earthquake magnitude.
• Being in an area where trenching is allowed by the government.

Deep seismic reflection data, which were available, show the extend of the deeper structure of the fault, between 500 and 2000 meters. Combination with field observation and aerial photographs (figure 2) enable the identification of the surface expression of the fault.

Figure 2: Aerial photo of the survey area (as indicated by the yellow square in figure 1). The Peel Boundary Fault is marked by the red dashed line and shows a scarp in the northeastern part of the photo. Lines A, B, and C in yellow represent the selected profiles for geophysical investigations.

The fault, denoted by the red dashed line, runs from the Northwest to the Southeast through the village of Neer. In the Northeast a clear fault scarp of 1 meter was observed in the field. While the fault scarp disappears towards the Southeast through agricultural activities.
Geophysical techniques were used to identify the fault near the surface. In this way, more confidence is obtained in the field observations. And, more importantly, the presence of structural phenomena around the fault can be resolved, if present. These structural phenomena could lead to the identification of paleo-earthquakes. Two techniques capable of deriving a high resolution image of the near surface are: ground penetrating radar (GPR) and resistivity measurements.
GPR measurements are based on differences in di-electric constant, of the medium. Reflections are mapped in the time domain, showing the interfaces where changes in di-electric constant occur. Figure 3 shows the results obtained for GPR measurements along line B in figure 2.

Figure 3: GPR section with 120 MHz antennas along line B

A raw estimation of depth can be obtained by multiplying the one-way travel time with 0.07 m/ns. It is obvious that at a distance of 185 meters, a difference in radar velocity structure appears. There is a increase in resolution at 185 meters going to larger distances. Structural phenomena, with respect to di-electric constant, show up at in the right hand side of the section. At 230 meters an other change in properties of the underground is visible, although less pronounced than at 185 meters.
Resistivity methods are based on measuring potential differences between two electrodes, induced by a known applied current. The electrodes are placed on top of the surface. A pseudo-section of average resistivities is obtained, through Ohm's law. Inverting the pseudo-section towards real resistivity values results in the electrical tomography section shown in figure 4. A sudden increase in resistivity appears at a distance of 175 meters. Also, an anomalous zone is present around 223 meters. The ground water table can be identified at 3 meters, since structures beneath 3 meters are smoothed due to the severely decreased penetration depth.

Figure 4: Electrical tomography section

Both sections show an anomaly at approximately 180 and 225 meters. The increase in resolution of GPR in the lower block can be explained by the higher resistivities found in the electrical tomography section. The upper block of the Peel Boundary fault is known to be much wetter (higher ground water table) than the lower block. Radar waves are extremely reduced by water, lowering the penetrating depth and resolution. This is confirmed by the lower resistivities in the upper block. A possible explanation can be that clays within the opened fault plane block the flow a water from the higher upper block towards the topographically lower hanging wall. Furthermore, a small basin like structure is found both on the radar and electrical tomography section between approximately 180 and 225 meters, giving confidence in possible structural phenomena.

Camelbeeck T. and Meghraoui M. (1998), Geological and geophysical evidence for large paleo-earthquakes with surface faulting in the Roer Graben (northwest Europe). Geophys. J. Int., 132, 347-362.