Geophysical study of Dune du Pilat Ly`es Amermoussa, Benoit Blanco, St´ephane Batany, Fr´ed´eric Chasserot, Sofia El Filali and Fr´ed´eric Vidry Institut de Physique du Globe de Paris, France November 25, 2011 Abstract We led a geophysic survey on the Dune du Pilat to characterize paleosols: localizatio n, depth and thickness. For this purpose, we have used four methods: the self-potenti al (SP) and electric methods, multi-offset radar and GPR. SP should give us the local ization but the current intensity on the paleosols surface was not significant compare to the rest of the signal and did not allow us to locate paleosol. Then electric method was used to create an underground image and to define paleosols. This method is ba sed on a measure of resistivity difference due to the various layers’s properties. Finall y, a combined multi-offset radar and GPR study has given us the velocity model that permits to characterize localization and depth of paleosols. We manage to see an ev olution of the paleosols especially for the third which waves. That confirm the fact it w as an old dune surface. We were also able to show, thanks to a carfully made GPR data processing, the ripple of a paleosol 15m in the subsurface.

1

Introduction

The Pilat dune is the greatest sand dune in Europe. It is situate in the south of the Arcachon bay. As all dunes, it is still moving and a such mass can have disastrous consequence on the environment as forest, building. . . To prevent the displacement is important to predict and counteract this phenomenon. That why we try to map the underground because understand the past move can help us to anticipate the future displacement. To reach our objective, we use four geophysics technics: self-potential (PS), electrical resistivity tomography (ERT), multi-offset radar and grounding penetrating radar (GPR). We investigate on the entire length of the dune: electical methode in the south, GPR in north and multi-offset radar in the middle (figure 1).

2

Geoelectrical prospecting

To characterize the paleosols we used three geoelectric methods : Schlumberger, Wenner and WennerSchlumberger. We focus on Wenner-Schlumberger methods because it combins the Wenner’s resolution quality with the Schlumberger’s depth investigation. The profile has a north-east, south-west direction and measures 112 meters (figure 1, EPS). The electrodes are spaced by one meters.

2.1

Figure 1: Localization of the different study sites

Generalities

To process the resitivity data, we used RES2DINV software with the following settings : The convergence 1

3 NE

SELF-POTENTIAL PROFILE Wenner-Schlumberger logarithm incertitude

2 P3

SW

Potential variation 25

Measurement points

20

15

10

Δ V (mV)

Sand

5

0

-5

-20

0

0

20

40

Paleosol 3

-15

-25

60

80

Wetland

Position (m)

100

120

140

160

180

22

44

time (ns)

Figure 3: Wenner-Schlumberger logarithm incertitude model block

Paleosol 2

P2 or Groundwater flow P3

Diatom layer

-10

66

88

110

limit setting is 2, this indicates that the inversion process has converged if the RMS error doesn’t change Figure 4: Top panel : Potential variations according below two per cent. We choose a vertical-to-horizontal position. Bottom panel : GPR profil flatness filter ratio of 0.5, indeed, we saw with the radar GPR that paleosols are elongated horizontally. And for those currents and the presence of paleosols ,thus we all methods we selected a robust inversion. have recorded a 180-meter-long profile of self-potential profile (PS profile) in three sections (the first begins at 2.2 Inversion results and interpretation see level) and in parallel we have determined the dune’s slope. Note that we have not included the topography. In The obtained measurements needed processing. fact, our geolectric profile is tranverse and sparsely afFirst we have merged the three sections of the PS profil fected by changes in altitude (the GPS vertical error to the same reference (point 0). Then, considering that bar is +/- 3 meters, whereas, the topography varies by with SP method, the first order interpretation gives the only one meter). topographic profile, we substrated from PS profile the The inversion shows a low resistivity layer at the topographic profile to obtain an approximated straight surface (low resistivity are represented by color from line . (Figure1, uper panel), blue to green). In fact, we saw that paleosol three During the campaign, we spotted the localization of outcrop at the profile location. the two paleosols . The analyze of the processed SP We can follow the paleosol three with this low resisiprofile (Figure1), showed the presence of peak corretivity layer near the surface until around 3 meters deep sponding to each paleosol but the intensities are not in the north-east part of the profile ??. The strong resignificantly different from the other signals. Thus no sistivity (represented by color from orange to purple) significant correlation between peak and paleosol has indicates a broad sand layer between 4 meters and 10 been observed . So it is not possible to situate the meters ??. From 10 meters we find a low resistivity paleosols with the SP method. layer, although, it is difficult to determine the structure at this depth. Indeed, around 11 meters, we are reaching the resolution limits. This layer may be the paleosol two, or, an groundwater flow above the pale- 3.1 GPR comparison osol two. Next , we have compared the corrected PS profile with The linear model block with a 300 ohm.meter con- the GPR profile recorded by the group 2(the figure 1 tour 2 shows the same results. lower panel ) ??, the paleosol’s localizations are spotted The logarithm incertitude model block 3 indicates in grey and we observed that each one are preceded that the inversion data are correct, despite an increase by a potential peak. The others potential peaks may of incertainty at the edges. be explain by the presence of “under-paleosols” as the According to inversion data, we can say that paleosol diatom layer (in green). three waves along the dune. Under the paleosol three, A special zone observed on the field was found on we find a sand layer with high resitivity followed by the the SP profile. Near the profile‘s top, some wetland paleosol two, or, an groudwater flow. induce by the seepage (in blue) have the most impor-

3

Self-potential profile

tant spontaneous potential. It can be explain by the proximity of the source.

To conclude on the PS investigation, it is obvious this method is not able to characterize paleosol. The In order to characterize the different paleosols present signal created by current on the paleosols surface is too in the Pilat dune we have tested the potential spontalow to be significant. Without the radar profile, the SP neous profiling or self-potentialmethod. Paleosols are signal interpretation was not possible. known to be waterproof, so the water flows induced by rain and loaded with ions Na+ and Cl- carried by sea Even if a correlation between the electric signal and air cross permeable sand and run on paleosol surface the underground (GPR profile) has been found, the creating an electric current. Our aim was to correlate analyze with only the PS method has not been possible

4

GROUND-PENETRATING RADAR

NE

3

Wenner-Schlumberger linear model block with a 300 ohm.m contour

P3

SW

Sand

P2 or Groundwater flow Low resistivity

Strong resistivity

Figure 2: Wenner-Schlumberger linear model block

Figure 6: Ripple of Paleosol 3 Paleosol 4

Figure 5: GPS positions of 12 GPR profiles

4 4.1

Ground-penetrating radar Presentation of the method

We began our geophysical prospection of La Dune du Pilat with a MALA Ground Penetrating Radar (GPR) acquisition. It is allowing a rapid collection of continuous subsurface data with maximum effectiveness in dry or fresh water saturated sediments with low clay content, making it an ideal imaging tool to investigate a dune. First, we used the GPR in mono-static mode with a 100Mhz antenna. An antenna transmits an electromagnetic pulse across different layers. When this wave encounters an interface, a part is scattered and a part is reflected. This reflected energy travels back to the surface where it is then recorded. The next step was to make a multi offset experience, with different offset between the emitting and the receiving antenna on the same profil.

Distance (in m)

Figure 7: Profile J, K and L L because they showed the best signal to noise ratio. Few profiles were done at the sea level to improve the horizontal resolution in this part where we expected to find the iron-rich/saline water contact. The data were post-processed using Reflexw software, including DeWoW, zero-time correction, static correction, gain function and a running average 2D filter. Then we made a topography correction thanks to the GPS data and Google earth. The processing sequence is detailed in 4.2.

As shown in figure 6, I profile is really interesting. It shows the ripple of paleosol 3 in the West – East direction and from 10m to 15m under the subsurface although it should be flat. This is the first time such a 4.2 Mono-static mode.100 MHz phenomenon is highlighted. We are sure that is not an error in the topography correction because we can see The acquisition was performed along W-E direction it going up besides the acquisition was uphill. We can with the 100 MHz MALA GPR. We made 12 profiles observe it with a lesser extent on profil J, K and L in of 250 m each as shown on figure 5: figure ??. It was a painful experience but the results are worFour layers are visible on figure 3. The first layer thy. We focused on the purple’s profiles: I, J,K and situated at 28 meters can be interpreted as the third

5

CONCLUSION

4

Figure 8: Profile J, K and L paleosol. The second and third layer, with a quite low reflectance, could correspond to an intermediate paleosol, composed of diatome. The elevation is respectively 15 and 10 meters. The fourth layer is located at the ledge of the beach. It certainly corresponds to paleosols 1 and 2, which are visible on the western side of the dune. On the beach, the low-quality signal is due to the water that is present few centimeters in the subsurface. Thanks to a combination of GPR and GPS data, we edit figure 8. This 3D view on Google earth enables us to validate our paleosol sensitivity. Indeed, we can appreciate the impact of the third paleosol, i.e. the gray line on the middle of the dune, on GPR data.

4.3

Multi-offest radar

ground

penetrating

Multi off-set GPR data were collected at the middle of the dune du pilat in the W-E direction along a 110m profile. There is 4 uphill and 4 downhill acquisitions. The off-set range going from 0.5m to 4m, increasing each 0.5m. We apply prestack processing techniques such as shown in figure 5: DeWoW, Static correction, a running average, and a Kirchhoff migration. Then we applied a topography correction thanks to the GPS data. Figure 10 shows all the downhill data after processing and topography correction. We can appreciate that the processing leads to a good repeatability between each offset. At this point, we were not able to stack these processed data with reflexw software.

5

Conclusion

To end this study, we can say that we managed to definite the paleosols. From all the technics use, the GPR was the most efficient and gives us the better results especially with the pseudo-3d block. The weak resistivity difference between paleosols and sand on one side and the multitude of little paleosols on the other hand compromise the efficient of the electric methods PS and ERT. The environment cleanliness gives us some GPR

Figure 9: Processing of multi-offset data

5

CONCLUSION

5

Figure 10: Downhill processed multi-offset data profiles easy to read and analyze. Unfortunately, we were unable to complete the processing of the multioffset GPR data because of our lack of knowledge in this field. However, the rippling of the third paleosol were clearly identified, giving us the kind of dune surface was during the year 1600, before the dune was consolidated. Thus, it was an entertaining field experience and an enlightening processing sequence thanks to a good communication and sharing of knowledge in the team.

Acknowledgments The team thanks all the supervisors V. All`egre, A. Maineult, G. Occhipinti and S. Roumejon, the other groups to their help and the share of result and the IPGP for supporting the fieldwork.