GEOPHYSICS,VOL.66,NO.1(JANUARY-FEBRUARY 2001);P .78–89,9FIGS.
Case History
The use of geophysical prospecting for imaging active faults in the Roer Graben,Belgium
Donat Demanet ∗,Fran¸c ois Renardy ∗,Kris Vanneste ∗∗,Denis Jongmans ‡,Thierry Camelbeeck ∗∗,and Mustapha Meghraoui §
ABSTRACT
As part of a paleoseismological investigation along the Bree fault scarp (western border of the Roer Graben),various geophysical methods [electrical profiling,elec-tromagnetic (EM)profiling,refraction seismic tests,elec-trical tomography,ground-penetrating radar (GPR),and high-resolution reflection seismic profiles]were used to locate and image an active fault zone in a depth range between a few decimeters to a few tens of meters.These geophysical investigations,in parallel with geomorpho-logical and geological analyses,helped in the decision to locate trench excavations exposing the fault surfaces.The results could then be checked with the observations in four trenches excavated across the scarp.Geophysical methods pointed out anomalies at all sites of the fault po-sition.The contrast of p
hysical properties (electrical re-sistivity and permittivity,seismic velocity)observed be-tween the two fault blocks is a result of a difference in
the lithology of the juxtaposed soil layers and of a change in the water table depth across the fault.Extremely fast techniques like electrical and EM profiling or seismic refraction profiles localized the fault position within an accuracy of a few meters.In a second step,more detailed methods (electrical tomography and GPR)more pre-cisely imaged the fault zone and revealed some struc-tures that were observed in the trenches.Finally,one high-resolution reflection seismic profile imaged the dis-placement of the fault at depths as large as 120m and filled the gap between classical seismic reflection profiles and the shallow geophysical techniques.Like all geo-physical surveys,the quality of the data is strongly de-pendent on the geologic environment and on the contrast of the physical properties between the juxtaposed for-mations.The combined use of various geophysical tech-niques is thus recommended for fault mapping,particu-larly for a preliminary investigation when the geological context is poorly defined.
INTRODUCTION
This paper describes the application of various geophysical prospecting techniques to locate and im
age Quaternary fault zones as part of a paleoseismological project (Meghraoui et al.,2000).Paleoseismology aims to determine the Late Pleistocene and Holocene history of near-surface faulting often associated with large earthquakes.This usually requires the excavation of shallow trenches across the trace of the suspected active fault.Active normal faults exposed at the surface are usually
Published on Geophysics Online July 11,2000.Manuscript received by the Editor March 22,1999;revised manuscript received June 15,2000.∗
Liege University,LGIH,Bat B19,4000Liege,Belgium.E-mail:ddemanet@ulg.ac.be.
‡Formerly Liege University,LGIH,Bat B19,4000Liege,Belgium.Presently LIRIGM,Universite Joseph Fourier-Grenoble 1,BP 53,F 38041Grenoble Cedex 9,France.E-mail:djongmans@ulg.ac.be.∗∗
Royal Observatory of Brussels,av.Circulaire 3,1180Brussels,Belgium.§CNR-CS,Geologia Tecnica,via Eudossiana 18,00184Rome,Italy.c  2001Society of Exploration Geophysicists.All rights reserved.
expressed in the topography as fault escarpments.However,in intraplate areas characterized by relatively low rates of tec-tonic deformation,the geomorphic expression of an active fault may be very
subtle as a result of the complex interplay among tectonic,depositional,and erosional processes or intensive agricultural exploitation.However,nondestructive geophysi-cal prospecting techniques may be applied to map the near-surface fault trace with great accuracy.In the last few years,a large number of high-resolution seismic reflection surveys have been conducted (e.g.,Williams et al.,1995;Palmer et al.,1997;
78
Imaging Active Faults79
Van Arsdale et al.,1998)to provide information on Quater-nary fault geometry and timing.For very shallow investigation, ground-penetrating radar(GPR),which can bridge the gap be-tween high-resolution seismic surveys and trenching,has been applied by Cai et al.(1996)in the San Francisco Bay region. At the border of Nevada and California,Shields et al.(1998) have used several geophysical techniques(seismic reflection, magnetics,and electromagnetics)to locate the extension of the Parhump Valley fault zone.This paper presents the results of a geophysical campaign performed in the Bree area(Roer Graben,northeast Belgium)as a reconnaissance tool prior to trenching,which included refraction seismic records,electro-magnetic(EM)and electrical profiling,elect
rical tomography, ground penetrating radar(GPR),and high-resolution seismic reflection profiles.The foremost aim of this investigation was to determine the exact position of an active fault to precisely locate a subsequent trench.A second objective was to image the fault zone at shallow depths,therby allowing a direct com-parison with trench data and hence a confident extrapolation of direct observations to greater depths.
GEOLOGICAL SETTING AND TECTONIC ACTIVITY The Roer Graben,which crosses three countries(Belgium, The Netherlands,and Germany),is bounded by two north–northwest,south–southeast-trending Quaternary normal fault systems(Figure1).The eastern boundary is defined by the Peel boundary fault,where the5.4-M W1992Roermond earthquake occurred(Camelbeeck and van Eck,1994);while the western boundary is defined by the Feldbiss fault zone,which is partially located in Belgium.Evidence of tectonic activity in the Roer Graben is given by(1)the strong subsidence during the last 150000years(Geluk et al.,1994),(2)the Quaternary faults and their associated morphology along theflanks of the graben, (3)the0.8–2-mm/yr vertical rate of deformation obtained by the comparison of levelings during the last100years(Van den Berg et al.,1994;M¨alzer et al.,1983),and(4)the present-day seismic activity(Camelbeeck and van Eck,1994).
小柯For the Feldbiss fault zone,tectonic activity is mainly indi-cated at depth by seismic profiles that show
more than600m of offset in Neogene deposits(Demyttenaere and Laga,1988)and about150m at the base of the Pliocene(De Batist and Versteeg, 1999).By considering the offset of the main terrace of the Mass River determined by Paulissen et al.(1985),Camelbeeck and Meghraoui(1998)obtain0.08±0.04mm/year for the average Late Pleistocene vertical deformation along the Feldbiss fault. Near the town of Bree(Figures1and2)and along the Feldbiss fault,a prominent northwest–southeast-trending fault scarp separates the Campine plateau to the west from the Roer Valley Graben to the east(Paulissen,1973).The geomorphic expression of the scarp consists of a10-km-long escarpment that has15–20m of vertical topographic relief(Figure2).The Belgian Geological Survey acquired150reflection seismic lines in the region with a dozen crossing the scarp(Demyttenaere, 1989).On different sections,the scarp coincides at the surface with the surface projection of the Feldbiss fault zone and can therefore be considered the morphological expression of the fault’s recent activity.
The Bree fault scarp corresponds to the northeastern border of the Campine Plateau(Figure2),which is covered by terrace gravels deposited by the Mass River(Zutendaal gravels)during the Cromerian(between770000and350000years BP)and which overly sands of Upper Miocene age(Diest Formation)(Paulissen et al.,1985).In the downthrown block (Roer Graben),the Zutendaal g
ravels have been eroded by the Rhine and Maas Rivers,which afterward deposited the Bocholt sands(Paulissen,1983).These formations constitute the basement on which the Maas formed its different terraces at the end of the Middle Pleistocene and during the Late Pleistocene.These terraces are the typical landscape of the region.The region was later covered with aeolian sands during the Saalian and Weichselian glacial ages,which were mixed with the other near-slope deposits in the vicinity of the fault scarp.Afinal phase of deposition created the Holocene alluvium in the center of the Maas Valley.The lithology logs of two boreholes(Van der Sluys,1997)drilled on each side of the scarp are given in Figure3.On the Campine plateau(hole H1),the Zutendaal gravels directly overlie the Upper Miocene sands of the Diest Formation,which were encountered at 11m depth.In the Roer Graben(hole H2),the thickness of the Middle Pleistocene river terraces reaches40m,while the top of the Diest Formation was found at233m depth,
below
F IG.1.Quaternary faults and seismic activity in the lower Rhine embayment.The Bree fault scarp is located along the Feldbiss fault southeast of the town of Bree.
80Demanet et al.
a succession of sand and clay layers from Lower Pliocene to Upper Pliocene.The depth difference in stratigraphic horizons between the two boreholes gives strong evidence of tectonic activity along the Feldbiss fault zone.
Multiple scarplets are superposed on the overall fault scarp,and the frontal fault trace consists of an en echelon geom-etry that suggests a component of left-lateral slip.The fault dips 70◦northeast and offsets young deposits (mainly late Weichselian aeolian cover sands and local alluvial terraces).Leveling profiles across the frontal fault scarp yield a vertical displacement ranging from 0.5to 3m.A three-year detailed paleoseismic investigation (1996–1998)shows that this frontal scarp corresponds to the latest coseismic (occurring during an earthquake)surface ruptures along this segment of the Feld-biss fault.These studies (Camelbeeck and Meghraoui,1996,1998)suggest that the most recent large earthquake occurred along the fault scarp between 610and 890A.D.and produced a vertical coseismic displacement of 0.5to 1.0m,with a mini-mum moment magnitude estimated as 6.3M W .Paleoseismic in-formation combining the trench and geomorphic observations suggests the occurrence of two surface-faulting earthquakes during the last 20000years.A third dates between 28000and 42000years BP .吻别歌词
DATA ACQUISITION
Figure 2shows the location of the four sites where geophysical profiles were performed perpendicular to the fault strike.At these sites trenches were later excavated for a paleo-seismic study (Meghraoui et al.,2000).Six geophysical methods were applied across the scarp:(1)electrical profiling,(2)EM profiling,(3)electrical tomography,(4)GPR,(5)
seismic
F I
G .2.Location and geological map showing the frontal escarpment of the Feldbiss fault near Bree and the studied sites (labeled 1to 4).The seismic line (SL)is located between sites 1and 2,at a right angle to the scarp.H1and H2are the boreholes described in Figure 3.Contours indicate
topography.
F I
G .3.Stratigraphic logs of boreholes H1and H2along a schematic southwest–northeast cross-section (after Van der Sluys,1997).
Imaging Active Faults81
refraction tests,and(6)high-resolution seismic reflection pro-files(Telford et al.,1990;Reynolds,1997).Geophysical tests were performed along the axis of each planned trench except for the seismic reflection profile,which was carried out between sites1and2(Figure2).As afirst step,the variation of the ap-parent ground resistivity along the scarp was measured with electrical and/or EM profiling.EM surveying was conducted with two separate coils connected by a reference c
able moved along the profile at discrete intervals with a constant coil spac-ing(Reynolds,1997).The instrument provides a direct reading of the apparent resistivity of the ground.In this study,the mea-surement spacing was5m and the intercoil separation was 10m.With horizontal coils,the maximum contribution to the secondary magneticfield theoretically arises from a depth of around4m.
In electrical profiling,a Schlumberger configuration with cur-rent electrodes spaced12m apart(50m for site4)was moved perpendicular to the profile,providing measurements of the apparent resistivity of the ground as a function of distance.An electrical tomography survey was performed using the Lund imaging system(Dahlin,1996)with a Wenner configuration and an electrode spacing of1or2m.The data were processed with the algorithm proposed by Loke and Barker(1996)to ob-tain a resistivity section.According to the profile length,the investigation depth was between5and15m.GPR profiles were also performed at three sites with a120-MHz transmitting an-tenna and at site3with a50-MHz antenna.A static correction was made with a mean velocity of80to90mm/ns determined from scattered events.The penetration depth strongly depends on the ground resistivity(ranging between50and500ohm-m in the Bree area)and was limited to a few meters.The GPR vertical resolution was smaller than0.5m with the120-MHz an-tenna used.At two sites,seismic refraction profiles,44and70m long,were carried out with a geophone spacing of1m and
three tofive shots.The source was a hammer,and twenty-four10-Hz geophones were connected to a16-bit seismograph.Finally, one seismic reflection line was run in a northeast–southwest direction perpendicular to the fault scarp(Figure2).The pro-file extends150m with a4-m source interval.A gun provided the source,stackedfive times for each source location.The op-timum window(Hunter et al.,1984)was determined from30to 56m from a walkaway test.Data were recorded with a16-bit seismograph from40-Hz geophones.The stacked data have a maximum of six-fold subsurface coverage.Processing was performed using SU software(Cohen and Stockwell,1998), and the sequence included static corrections,F-Kfiltering, NMO corrections,prestack band-passfiltering,CDP stack and poststack band-passfiltering.
RESULTS AND INTERPRETATION
The results of geophysical tests parallel to trenches T1to T4 are presented in Figure4and Figures6to8as well as a simpli-fied geological description of each trench.The seismic reflec-tion profile is shown in Figure5.
Site1
Thefirst site is located near a stream that cuts a small uplifted alluvial terrace.The trench,which is onl
y2m deep,reveals late Weichselian cover sands,the upper part of which has been reworked by the small river(Figure4a).Disruption of(1)two
gravel horizons within the cover sands and(2)the bleached
Holocene soil at the top indicates the near-surface presence
of a normal fault dipping to the northeast and closely aligning
with the frontal escarpment.An overlying soil bed just below
the plough zone does not appear to be affected.
Electrical profiling data clearly delineate the fault at a dis-
tance between50and65m(Figure4a)by a sharp increase of
the apparent resistivity values,from70ohm-m in the south-
west block to more than250ohm-m in the northeast block
(Figure4b).An accurate location(within a few meters)of
the fault is,however,impossible to assess.The electrical to-
斗罗大陆叶知秋mography section(Figure4c)shows a strong lateral resistivity
variation around50m with a contact dipping to the northeast.
In the southwest block,a2-m-thick resistive layer overlies a
conductive formation,while the northeast block consists only
of the resistive layer.Here,the fault juxtaposing different soil
layers can be located at the surface with an accuracy<2m.
A second strong lateral resistivity variation at20m could be
江南style中文版interpreted as a fault dipping to the southwest.However,this
was shown neither on the seismic profile nor in the trench,and
the anomaly probably results from a sedimentary variation.
A70-m-long seismic refraction profile was performed across回音哥照片
the scarp.The time–distance graph inferred from the refracted
wave analysis for the direct shot(Figure4b)shows an unusual
decrease of the apparent velocity from1690to720m/s in the
subsurface.This crossover point is located around50m and
fits perfectly with the position of the fault.The interpretation
of the seismic data(Figure4a)with the generalized recipro-
cal method(Palmer,1981)shows that the conductive underly-
ing layer is characterized by a relatively high seismic velocity
(V p=1400m/s).In the southwest part of the section,this hori-zon is covered by a thin layer with a velocity of470m/s,which
dramatically increases in depth across the fault to reach4m in
the hanging wall.The limit between the two seismic horizons
could correspond to the depth of the water table,which was
less than2m in the footwall.Both geophysical methods clearly
indicate the presence of a fault below the topographic scarp,
juxtaposing two blocks with different resistivity and seismic
velocity values.The corresponding GPR section is presented
in Figure4d,where thefirst30ns corresponding to the direct
wave have been muted.The maximum penetration depth is
about4m,corresponding to a two-way traveltime of100ns.
In the southwest part,the section reveals two main horizontal
reflectors(R1and R2),which are clearlyflexured and cut by
荒漠甘泉歌曲two fault branches.The main one(F1)is located at about50m
along the profile,whereas the second fault branch F2prob-
ably does not extend to within the reach of the trench.The
shallower reflector(R1)is located at1.6m depth(40ns)and
correlates with the lower gravel horizon exposed at the bottom
of trench1.The northeast part of the section is characterized by
a wedge shape with a southwest-dipping strong reflector(R3)
at its base.The base of the wedge is located at3.2m depth.The
different layers inside the wedge appear to beflexured in the
vicinity of the fault.
Seismic line SL(Figure5),150m long,trends southwest–
northeast and crosses the frontal escarpment(F)at a right
angle.In the Roer graben(northeastern block),the seismic
section reveals several well-defined reflections down to0.2s.
These seismic horizons are cut at105m by a fault(F)whose
82Demanet et al.
F IG.4.Site1.(a)Schematic stratigraphic cross-section and seismic velocity model.(b)Electrical profiling(EP)and seismic
traveltime curves(SP).(c)Electrical tomography.(d)Radar section(120MHz).