Skip to navigation – Site map

HomeNumérosvol. 20 - n° 2Piezometric thresholds for trigge...

Piezometric thresholds for triggering landslides along the Normandy coast, France

Seuils piézométriques pour le déclenchement de glissements de terrain sur les versants côtiers normands, France
Candide Lissak, Olivier Maquaire, Robert Davidson and Jean-Philippe Malet
p. 145-158

Abstracts

The Cirque des Graves landslide in Normandy (N-W France) is a slow-moving and deep-seated coastal landslide characterized by surface displacement of a few centimeters per year (1-10 cm/yr) and regularly affected by brutal accelerations with several meters of displacements (January 1982, February 1988, March 1995 and March 2001). For these reasons, and because of the presence of economical and physical stakes, the landslide has been progressively monitored since 1985. Previous researches demonstrated qualitatively that these accelerations were controlled by the slope hydrology. To improve the knowledge on the slope dynamics, a combination of historical data and slope instrumentation is necessary. The actual monitoring system is based on twenty-four cemented benchmarks, three permanent GNSS receivers and several hydro-meteorological observation points. The aim of this study is to associate groundwater levels to the landslide kinematics to define critical piezometric thresholds to trigger landsliding. Thresholds have been identified using long-term continuous single point time series and shorter, but more spatially distributed, piezometric information. Empirical piezometric thresholds for the landslide triggering are defined taking into account the antecedent hydro-climatic conditions. Two main situations were identified: (1) major landsliding events associated to long-lasting rainfall periods (several month) with groundwater water elevation of more than two meters above the mean annual level observed for the period 1976-2013 (return period of over 6/7 years); and (2) moderate landsliding events associated to low intensity rainfall periods with a limited groundwater rise responsible for a moderate seasonal kinematics (return period of every years).

Top of page

Editor's notes

Article soumis le 3 mai 2013, accepté le 30 janvier 2014.

Full text

This research was funded through the RiskNat project ‘SISCA: Système intégré de Surveillance de Crises de glissements de terrain argileux’ (2009-2013) of the Agence Nationale de la Recherche (ANR-08-RISK-0009) and the Project CPER GR2TC: Gestion des Ressources, Risques et Technologie du domaine Côtier’ (2007-2013). The permanent GNSS system and the automated processing of the observations have been setup by the OMIV-EOST Unit at University of Strasbourg as part of the French Observatory on Landslide. We would also like to thank the LETG-Caen Geophen laboratory team for helping us to install the device.

Introduction

1Traditionally, prediction of landslides is based on historical data analysis (Malet et al., 2003a; Guzzetti et al., 2004; Floris and Bozzano, 2008) and field investigations (Hansen, 1984; Soeters and van Westen, 1996; Malet et al., 2003b) to characterize the hydro-climatic conditions operating on the slope stability and the landslide kinematics. The underlying idea is to determine rainfall (duration / intensity) and piezometric thresholds above which first-time landslides are triggered or acceleration are observed on continuously active landslides (Caine, 1980; Crozier, 1986; Corominas et al., 2000; Iverson, 2000; Aleotti, 2004; Gonzalez-Garcia and Mayorga-Marquez, 2004; Guzzetti et al., 2004; Guzzetti et al., 2008).

2Several studies have demonstrated the control of rainfall inputs (van Asch et al., 1999; Delmonaco and Margottini, 2004) and groundwater fluctuations (Bogaard et al., 2000; Cappa et al., 2004) on the observed amplitudes of displacement.

3Along the Normandy coast, several landslides are active since decades, and their dynamics is controlled by slope hydrology (Maquaire, 2000). First investigations at the Cirque des Graves landslide highlighted that this landslide is characterized by a continuous sliding (several centimeters of displacement per year), but also by alternating periods of acceleration (several decimeters or meters of displacement per year) and deceleration.

4A first geodetic network, consisting in topographic benchmarks, was initiated in 1985 (Maquaire, 1990) and has been used to record acceleration periods in February 1988, March 1995 and March 2001. Some of these benchmarks are still on-site and ensure the long-term continuity of observations. To improve the knowledge on the landslide dynamics and to specify the spatial distribution and the temporal evolution of the velocity, high accuracy measurements started in 2008 using 1) repeated measurement campaigns of new benchmarks with GNSS implanted in different landslide units, and 2) permanent dual-frequency GNSS of the French Landslide Observatory – OMIV (eost.unistra.fr/omiv). Three GNSS stations have been implanted to provide high frequency observations with hourly solutions of positioning.

5The geodetic network is completed with a hydro-meteorological network (Lissak et al., 2010) consisting in multiple piezometric sensors and meteorologic stations also integrating observations from regional databases of groundwater levels and historical rainfall data from the French Meteorological Survey.

6The recent observations confirm the permanent activity of the landslide with displacement rates of a few centimeters per year. The acquisition of high frequency GNSS data provides evidence of spatial differences in the kinematic pattern of sub-units within the landslides, and highlighted the seasonal behavior of the landslide with higher displacement rates observed in spring and in autumn.

7In order to define some indicators of slope acceleration, the objective of this work is to identify possible hydro-meteorological thresholds associated to changes in the kinematic pattern.

Study area, material and methods

Study area

8Along the west margin of the sedimentary Paris Basin (north of France), at the edges of the Pays d'Auge plateau (fig. 1), several active landslides between Trouville and Honfleur occur (Cruden and Varnes, 1996).

Fig. 1 – Location and geomorphological setting of the Cirque des Graves landslide.
Fig. 1 – Localisation et contexte géomorphologique du glissement du Cirque des Graves.

Fig. 1 – Location and geomorphological setting of the Cirque des Graves landslide. Fig. 1 – Localisation et contexte géomorphologique du glissement du Cirque des Graves.

A: topographic context with location of the active landslide along the coast between Trouville and Honfleur. B: 3D view of the Cirque des Graves landslide in 2006. C: geological cross-section. 1: altitude; 2: coastal slope; 3: Pays d’Auge plateau; 4: active coastal landslide; 5: road D513; 6: buildings; 7: surficial deposit; 8: chalk; 9: glauconitic sand; 10: marls; 11: sandstone; 12: spring; 13: sliding surface.
A : contexte topographique régional et localisation des glissements de terrain actifs le long de la côte entre Trouville et Honfleur. B : vue 3D du glissement du Cirque des Graves en 2006. C : coupe géologique schématique. 1 : point côté ; 2 : versant côtier ; 3 : plateau du Pays d’Auge ; 4 : glissement de terrain actif ; 5 : route départementale 513 ; 6 : bâti ; 7 : formations superficielles ; 8 : craie ; 9 : sable glauconieux ; 10 : marne ; 11 : grès ; 12 : résurgence ; 13 : surface de glissement.

9These landslides are located on low-altitude convex-concave slopes of the Pays d’Auge Plateau (maximum elevation of 140 m above sea level) and present a complex morphology of composite rotational slides. This work focuses on the Cirque des Graves landslide (municipality of Villerville). The Cirque des Graves landslide is the most active and the largest roto-translational landslide of the region with a surface of 47 hectares. As outlined in the figure 1c, the lithostratigraphic scheme consists of Jurassic sedimentary rocks with superimposed strata of almost 10 m thick sandstone plunging gently to the South-East (10-20%), marls (25 m in thickness) and sand overlaid with chalk material. On the plateau, the chalk can exceed 50 m in thickness and is covered with flint clay, slope deposits whose thickness is very variable.

10The slope destabilization started probably during the Holocene period (Flageollet and Helluin, 1987) as outlined by the presence of large and high scarps (10-15 m height) before the main acceleration of January 1982 (fig. 1A). Three other major accelerations (February 1988, March 1995 and March 2001) have been observed and characterized by several meters of displacement creating high scarps by retrogression and large cracks (fig. 1B). The four accelerations have induced several damages to buildings and road infrastructures (fig. 1). The resulting topography is representative of successive rotational slides (fig. 1C) with typical morphological features such as opened fissures, fresh scarps of various sizes, depressions, counter-slopes and lobes at the toe. Independently of these acceleration events, the landslide is characterized by yearly average surface displacement rates of 5-10 cm/yr.

Geodetic survey network

11The actual network is composed of 24 cemented benchmarks extending over 1 km east to west. Some of them are still from the original network implanted in 1985. Two benchmarks are located on the stable parts of the slope and are used as control points to assess the quality of the measurements (Gili et al., 2000). The benchmarks have been implanted according to the accessibility, the vegetation constraints and the landslide detailed morphology (Lissak et al., 2014).

12The entire landslide body cannot be investigated because of the dense vegetation cover in the Western unit (the vegetation prevents any direct visual observations and the GNSS signal is not receivable), and large displacements (shallow landslide) at the toe. So these sectors are not integrated in the monitoring network due to the impossibility to anchor benchmark in thick muddy material.

13Until 2008, the position of the benchmarks has been measured by geodetic campaigns using a total station, and after 2008, through campaigns of differential GNSS. GNSS techniques have been increasingly employed to monitor landslide movements (Gili et al., 2000; Malet et al., 2000; Malet et al., 2002a; Squarzoni et al., 2005; Peyret et al., 2008) and vertical movements of unstable slopes (Baldi et al., 2008). For the study case, the acquisition protocol consists of repeated GNSS observations of the benchmarks of several minutes to several hours. Various studies have shown that the precision of the measurement in the three dimensions evolves according to the duration of the session (Malet et al., 2002a; Brückl et al., 2006; Peyret et al., 2008; Zhu et al., 2012). Several tests (Dixon, 1991; Malet et al., 2002b) have been conducted to define an optimal session duration of 17 minutes to obtain a horizontal accuracy of 0.7 cm in the East and North coordinates, and of 1.3 cm in the Up coordinate. The observations from the GNSS campaigns have been processed with the TGO software developed by Trimble©.

14Since July 2009, high frequency observations are acquired with three permanent GNSS stations to record very low amplitude displacements (fig. 2 and fig. 3).

Fig. 2 – Monitoring network for the Cirque des Graves landslide.
Fig. 2 – Réseau de surveillance géodésique du Cirque des Graves.

Fig. 2 – Monitoring network for the Cirque des Graves landslide. Fig. 2 – Réseau de surveillance géodésique du Cirque des Graves.

A: Cirque des Graves landslide. B: Pays d’Auge plateau. 1: piezometer; 2: wells; 3: inclinometer; 4: piezometer with permanent sensor; 5: geobead probe; 6: meteorological station; 7: topographic benchmark; 8: permanent GNSS receivers; 9: active landslide; 10: Cirque des Graves landslide.
A : Cirque des Graves. B : plateau du Pays d’Auge. 1 : piézomètre ; 2 : puits ; 3 : inclinomètre ; 4 : piézomètre avec capteur permanent ; 5 : sondes Geobead ; 6 : station météorologique ; 7 : repère topographique ; 8 : récepteurs GNSS permanents ; 9 : glissement de terrain actif ; 10 : glissement du Cirque des Graves.

Fig. 3 – Annual average displacement rate at Cirque des Graves landslide between 2 April 2008 and 25 April 2012.
Fig. 3 – Vitesses moyennes annuelles des déplacements mesurés au Cirque des Graves entre le 2 avril 2008 et le 25 avril 2012.

Fig. 3 – Annual average displacement rate at Cirque des Graves landslide between 2 April 2008 and 25 April 2012. Fig. 3 – Vitesses moyennes annuelles des déplacements mesurés au Cirque des Graves entre le 2 avril 2008 et le 25 avril 2012.

1: average annual velocity; 2: topographic benchmark; 3: permanent GNSS receiver; 4: limit of the active zone; 5: coast line; 6: chalk panel limit; 7: location of the schematic profile of fig. 4.
1 : vitesse moyenne annuelle ; 2 : repère topographique ; 3 : récepteur GNSS permanent ; 4 : limite de la zone active ; 5 : trait de côte ; 6 : limite de panneau de craie ; 7 : localisation du profil schématique de la fig. 4.

15Each GNSS station consists of a GNSS antenna TRM41249.00 (with radome TZGD) and a dual-frequency receiver Trimble© NetRS. One GNSS station has been set up on the stable part of the slope (VLRV). Two GPS receivers (VLRH and VLRB) are located on the active part of the landslide, on two subsequent compartments. The baseline between the receivers VLRV and VLRH is 465 m and between VLRV and VLRB 480 m. The acquisition frequency is 30 s over 24 h sessions. An automatic solution and an internet connection allow a near real time (j + 2 days) data acquisition (Malet et al., 2013). This network is included into the GPS Data Service of the French Landslide Observatory OMIV. GPS raw data and code phase (Malet et al., 2002a; 2002b) are calculated with the GAMIT / GLOBK software developed by the MIT (Massachusetts Institute of Technology). The automatic daily routine developed by OMIV-EOST Strasbourg takes into account the different tropospheric and ionospheric parameters for a 3D position accuracy of 0.5 cm. The procedure is presented in Malet et al. (2013).

Hydro-meteorological survey network

16The hydro-meteorological observation network consists of a permanent survey of ground water levels in boreholes, pore water pressures and rainfall at several locations (fig. 2). The piezometric network comprises twenty-nine observation points including four wells, seven boreholes equipped with inclinometer tubes and eighteen piezometers. The depth of the piezometers ranges from 5.4 to 29.0 m. The groundwater levels are measured manually except for six piezometers equipped with automatic probes. These piezometers have been selected according to their location, the amplitude of the groundwater level changes and are considered as representative of the different landslide units.

17In addition, to analyse the groundwater behaviour in the stable chalk formation overlaying the landslide, piezometric observations at the Danestal borehole are used. This piezometer, located 15 km away from the study site, has been selected among the regional BRGM groundwater monitoring network. Monthly observations for the period 1974-2011 and daily observations since 2011 are available.

18Pore water pressures and soil humidity are also monitored in depth with four multi-parametric probe Geobeads designed by the Alert Solutions (Peters et al., 2010). Four probes have been set up in the borehole C3 (fig. 2) at several depths (Cell 1: -1.0 m, Cell 2: -2.0 m, Cell 3: -4.00 m, Cell 4: -5.80 m) close to the permanent GNSS receiver VLRH, two continuous logging piezometers and one inclinometer. Daily pore water pressure, temperature and inclination data are available using online connexion with the station.

19The meteorological network consists of two meteorological stations. One is located on the plateau, 4 km from the landslide (altitude 144 m), at Saint-Gatien; this station from the French Meteorological Survey (Meteo-France) provides daily data of temperature, evapotranspiration, rainfall and effective rainfall since 1949 (Barry, 1981). Near the landslide, a meteorological station has been implanted in 2013, at Villerville (fig. 2).

Kinematic pattern of the landslide for the period 1985-2012

20Analysis of the geodetic data is conducted to characterize the spatial distribution and temporal evolution of the landslide at two time scales. The recent observations of GNSS campaigns and permanent monitoring are analysed for the period April 2008-April 2012; the oldest observations of geodetic benchmarks monitored by tacheometer are analysed for the period 1985-2008.

Spatial pattern of the surface displacements

21The spatial distribution of the annual average of displacements is presented on figure 3. Between April 2008 and April 2012, the measurements of the benchmarks show a N-NW general trajectory. During four years, the cumulative displacements range between 4 cm (n°43) and 46 cm (n°404) (standard deviation of 10 cm) with an important spatial variability per unit.

22The eastern part of the landslide is characterised by a cumulative displacement of about 7-8 cm at benchmarks 163 and 164. The benchmarks 203, 68 and 58, located at the lower part, are affected by cumulative displacements between 16 and 19 cm. In the same area, the same order of magnitude was estimated by the two GNSS receivers with values ranging from 5 cm (VLRH) to 9 cm (VLRB) since 2009. On the eastern part, displacement values decrease between 4-6 cm and 10 cm at the benchmark 450. In the central part of the landslide, just above the area affected by shallow landslides, cumulative displacements are between 18-25 cm (401, 402 and 405) and 30 to 46 cm (respectively the 403 and 404 benchmark). Downhill, the toe of the landslide is characterised by shallow landslides and mudflows with large displacements estimated at more than 2 m in four years.

23The landslide consists of several morpho-structural units (or compartments) characterized by the presence of chalk blocks or panels (Lissak et al., 2014). It has been demonstrated that the displacement measurement acquired on a specific point implanted on a panel can reflect the global movement of the entire chalk panel or compartment. Thus, all compartments are affected by different displacement rates. To define the spatial distribution of the annual average displacement rate, the point observations are extrapolated (fig. 3 and fig. 4) considering the slope morphology and internal structure.

Fig. 4 – Uphill and downhill distribution of landslide displacement rate (profile location in fig. 5).
Fig. 4 – Répartition amont-aval des vitesses moyennes des déplacements de surface selon une coupe schématique (tracé de la coupe sur la fig. 5).

Fig. 4 – Uphill and downhill distribution of landslide displacement rate (profile location in fig. 5). Fig. 4 – Répartition amont-aval des vitesses moyennes des déplacements de surface selon une coupe schématique (tracé de la coupe sur la fig. 5).

1: average annual velocity; 2: global displacement vector; 3: vertical displacement vector; 4: horizontal displacement vector; 5: sliding surface; 6: permanent GNSS receiver.
1 : vitesse moyenne annuelle ; 2 : vecteur de déplacement global ; 3 : vecteur de déplacement vertical ; 4 : vecteur de déplacement horizontal ; 5 : surface de glissement ; 6 : récepteur GNSS permanent.

24The spatial distribution of displacements presents a concentric organization around a central medial nucleus, corresponding to the area affected by shallow landslides and mudflows. This area is characterized by a kinematics estimated at more than 50 cm per year. Beyond the medial nucleus, the displacement rate varies between 7 and 12 cm per year (i.e. point 404 with more than 11 cm per year). Then, a large band including velocities between 7 and 1.5 cm per year follows. Finally, the uphill part, near the main scarp is affected by very low displacement rates with values close to 1.5 cm per year. This distribution of velocity corresponds to the model proposed by Maquaire (1990).

25Without the possibility to implement benchmarks in all the landslide units, detailed field observations on the buildings and the roads, allow revealing indicators of recent activity. For example, the presence of evolving fissures on the Road 513, at the central part of the landslide, that testifies of the progressive subsidence of the road. This highlights the importance of the vertical component of the displacement on the uphill unit of the landslide (fig. 4), near the main scarp.

Temporal pattern of the surface displacements

26The displacement observations since 1985, at the eastern unit of the landslide, recorded the displacement rates before, during and after the three major accelerations of 1988, 1995 and 2001.

27As can be observed in figure 5, the cumulative displacement curves confirm a significant variability of the displacement rates.

Fig. 5 – Cumulative displacements since 1985 of benchmarks located on the eastern part of the Cirque des Graves landslide.
Fig. 5 – Déplacements cumulés mesurés depuis 1985 des repères situés dans le secteur est du Cirque des Graves.

Fig. 5 – Cumulative displacements since 1985 of benchmarks located on the eastern part of the Cirque des Graves landslide. Fig. 5 – Déplacements cumulés mesurés depuis 1985 des repères situés dans le secteur est du Cirque des Graves.

28The total displacements are comprised between 8.20 m, at the landslide toe (n°68) and 0.90 m at the edge (n°63). The observations confirm a regressive evolution of the landslide with values decreasing from downhill to uphill. The three major crises are characterized by displacements comprised between several decimetres to several meters. Out of these acceleration periods, a permanent creeping of several centimetres of displacement per year is observed.

29The analysis of the permanent GNSS VLRH and VLRB baselines since August 2009 (fig. 6) confirm this statement and can be used as indicator of the landslide’s behaviour in time.

Fig. 6 – VLRH-VLRV and VLRB-VLRV baselines evolution between August 2009 and April 2013.
Fig. 6 – Longueurs des lignes de base VLRH-VLRV et VLRB-VLRV enregistrées entre août 2009 et avril 2013.

Fig. 6 – VLRH-VLRV and VLRB-VLRV baselines evolution between August 2009 and April 2013.Fig. 6 – Longueurs des lignes de base VLRH-VLRV et VLRB-VLRV enregistrées entre août 2009 et avril 2013.

30Between 1 August 2009 and 31 March 2013, according to coordinates, the station VLRB moved of ca. 22 cm and the VLRH station of ca. 8 cm. It should be noted that the evolution of the baselines is discontinuous, interrupted by four small acceleration phases separated by "stabilization" phases. We can notice a seasonal activity of a few millimetres per day or a few centimetres per week during the most active phases in the winter seasons. The amplitude of the displacement varies between the stations and the season. This highlights a slightly different kinematics between the two GNSS stations. The first phase of acceleration episodes triggered on 11 December 2009 and progressively decreased in May 2010 until the “stabilization” phase.

31While the behaviour of the two GNSS stations seems similar, a time lag for the onset of the acceleration is observed and the acceleration always starts earlier at the VRLB station. Delays between stations vary between 1 and 4 days for the different phases but could also be more important. It was the case during the acceleration phase observed in November 2010 with 20 days of delay. Outside the episodes of acceleration, the two GNSS did not record significant displacements but a "stabilization" phase with a very low progression of the stations in several months associated to the data error (2 mm in 192 days between June and November 2010, 146 days between in June and December 2010).

Hydro-meteorological pattern of the landslide

32The chalk aquifer is the main aquifer of the landslide area but the groundwater flows through the other geological layers such as surficial deposits and glauconitic sands. Thereby, according to the available piezometers, the annual average groundwater level can be different according to the landslide unit and ranges from 1.40 to 13 m depth. Another difference exists with the seasonal amplitude of the groundwater levels, ranging from several decimetres to 3 metres throughout the landslide area (Lissak et al., 2009; Devi Maharjan, 2011; Bogaard et al., 2011).

33The seasonal trend of the groundwater is clearly visible with a small variation of water level between dry and wet periods. Even if the seasonal amplitudes of the piezometers are heterogeneous, the periods of recharge are quite similar with a time lag between the piezometers in the beginning of the groundwater recharge (two days of time lag recorded between SD4 and C2). The maximum water level is reached from February to April, after winter recharge, and the minimum from August to October. This seasonality of the groundwater system observed can also be observed at the Danestal piezometer, on the plateau. The Danestal piezometer highlights a similar seasonal fluctuation of the groundwater level (fig. 8). We can notice three major water recharge periods in 1977-1989, 1992 -1996 and 1998-2002. These multi-annual periods of recharge are followed by periods of drainage of 3 to 4 years.

34The high groundwater levels are associated to the cyclic variations of the annual rainfall amount characterized by rainfall excess to the mean annual rainfall observed for the period 1949-2013. For this period, the annual precipitation is ranging from 500 (year 1953) to 1160 mm/yr (year 1974) and the average is 913 mm/yr (period 1981-2010). The maximum daily rainfall was in August 2003 with 72 mm. The maximum daily rainfall (> 20 mm/day) mostly occurs between October and March.

Relationships slope hydrology – slope kinematics

Rainfall – groundwater levels relationships

35To quantify the response of the groundwater fluctuation to the precipitations (fig. 7), hydrologic time series were analysed. Also, a statistical index can be defined to highlight the importance of the effective cumulative rainfall on the rise of the groundwater level.

Fig. 7 – Rainfall – groundwater levels (at SD4) relationships during a typical recharge period.
Fig. 7 – Relation pluie-nappe (à SD4) en période de recharge.

Fig. 7 – Rainfall – groundwater levels (at SD4) relationships during a typical recharge period. Fig. 7 – Relation pluie-nappe (à SD4) en période de recharge.

Fig. 8 – Relation between groundwater level variations at Danestal piezometer and effective rainfall between at Saint-Gatien between 1974 and 2013.
Fig. 8 – Relations entre les variations de la nappe de Danestal et les pluies efficaces annuelles entre 1974 et 2013.

Fig. 8 – Relation between groundwater level variations at Danestal piezometer and effective rainfall between at Saint-Gatien between 1974 and 2013. Fig. 8 – Relations entre les variations de la nappe de Danestal et les pluies efficaces annuelles entre 1974 et 2013.

A: groundwater level at Danestal piezometer since 1974. B: excessive rainfall compared to the mean annual effective rainfall at Saint-Gatien station.
A : fluctuation de la nappe enregistrée au piézomètre Danestal depuis 1974. B : excédent pluviométrique par rapport à la moyenne annuelle des précipitations efficaces enregistrées à la station de Saint-Gatien.

36The relation between rainfall and groundwater is more significant at the seasonal scale (fig. 9) with a progressive groundwater rise linked to successive rainfall events.

Fig. 9 – Relations between the displacements (GNSS VLRH and VLRB), piezometric levels (SD4 and C2), pore pressure (A4) recorded at Cirque des Graves and daily rainfall of Saint-Gatien station.
Fig. 9 – Relations entre les déplacements GNSS (VLRH et VLRB), les niveaux piézométriques (SD4 et C2), les pressions interstitielles (A4) enregistrés au Cirque des Graves et les précipitations journalières de la station de Saint-Gatien.

Fig. 9 – Relations between the displacements (GNSS VLRH and VLRB), piezometric levels (SD4 and C2), pore pressure (A4) recorded at Cirque des Graves and daily rainfall of Saint-Gatien station. Fig. 9 – Relations entre les déplacements GNSS (VLRH et VLRB), les niveaux piézométriques (SD4 et C2), les pressions interstitielles (A4) enregistrés au Cirque des Graves et les précipitations journalières de la station de Saint-Gatien.

37But in winter, after at least three consecutive rainy days, the groundwater level response (to effective rainfall) was estimated between 2 and 5 days (Lissak et al., 2009). The influence of cumulative precipitation events can be observed several days after a rainfall event. The groundwater recharge starts to become effective with a rainfall amount superior to 6 mm per day and an isolated rainfall event has no influence on the groundwater recharge.

Groundwater levels - kinematics relationships

38Only four landslide accelerations are known. These accelerations occurred during unusually wet periods characterised by a succession of heavy rainfall years (according to the mean annual effective rainfall). As shown in figure 8, the acceleration of January 1982, considered as the largest acceleration of the landslide, occurred during an excessive hydrological year (July 1981 - June 1982) with particularly abundant winter rainfall amount. The consequence of these contributions is an elevation at Danestal of the groundwater level nearly 1.50 m a few weeks before the landslide acceleration, following also a series of six consecutive years of excess rainfall (fig. 8). The accelerations of 1995 and 2001 occurred after two or three years of hydrological excess, and are associated to a significant rise of the groundwater level caused by this succession of very wet months. According to witnesses, the landslide activity progressively increased with small accelerations of a few centimetres per week related to the gradual rise of the groundwater level a few months before the acceleration.

39Since 2009 and the installation of the permanent GNSS receivers, no major acceleration occurred; only a few minor accelerations of a few centimetre displacements per year can be analysed. This can be explained by the low groundwater levels measured at Danestal (between -12.5 m and -14.0 m depth) linked to a period of annual effective rainfall amount below or slightly above the mean annual (435 mm at Saint-Gatien, fig. 8). A correlation between the seasonal trends of the rainfall at Saint-Gatien, the local and Danestal groundwater levels and the landslide velocity is however observed (fig. 9). The time lag between the groundwater rise and the acceleration beginning is ranging between one and four days.

Proposition of piezometric thresholds

40Hydrological thresholds are determined by empirical analysis of rainfall, groundwater fluctuations, and the surface displacements (fig. 9, fig. 10 and fig. 11).

Fig. 10 – Relations between piezometric fluctuations (SD4), surface displacements (VLRH and VLRB) and cumulative effective rainfall during the first period of acceleration (20 January 2010 to 13 May 2010).
Fig. 10 – Relations entre les fluctuations piézométriques (SD4), les déplacements de surface (VLRH et VLRB) et la pluie efficace cumulée durant la 1ère période d'accélération (20 janvier 2010 au 13 mai 2010).

Fig. 10 – Relations between piezometric fluctuations (SD4), surface displacements (VLRH and VLRB) and cumulative effective rainfall during the first period of acceleration (20 January 2010 to 13 May 2010).Fig. 10 – Relations entre les fluctuations piézométriques (SD4), les déplacements de surface (VLRH et VLRB) et la pluie efficace cumulée durant la 1ère période d'accélération (20 janvier 2010 au 13 mai 2010).

Fig. 11 – Relations between piezometric fluctuations (SD4), surface displacements (VLRH and VLRB) and cumulative effective rainfall during the fourth period of acceleration started in early December 2012.
Fig. 11 – Relations entre les fluctuations piézométriques (SD4), les déplacements de surface (VLRH et VLRB) et la pluie efficace cumulée durant la quatrième période d'accélération démarrée en début décembre 2012.

Fig. 11 – Relations between piezometric fluctuations (SD4), surface displacements (VLRH and VLRB) and cumulative effective rainfall during the fourth period of acceleration started in early December 2012.Fig. 11 – Relations entre les fluctuations piézométriques (SD4), les déplacements de surface (VLRH et VLRB) et la pluie efficace cumulée durant la quatrième période d'accélération démarrée en début décembre 2012.

41However, the number of significant events, since the network implementation, is too small to determine critical thresholds based on a statistical approach (van Asch and Buma, 1996; van Asch et al., 1999). To define a local threshold, the time series of piezometers SD4 and C2, located on the eastern part of the landslide close to the GNSS VLRB and VRLH are used. Both piezometers have been selected because of their location and their significant seasonal trends. To complete our analysis, the groundwater level was calculated from the pore water pressures measured by the Geobead probe (C3-A4) installed at a depth of 5.80 m close to C2 and VLRB.

42The seasonal behavior of the landslide has been demonstrated and associated to the periods of groundwater recharge. For the first acceleration (20 January 2010 to 13 May 2010), a preliminary threshold, can be defined at the piezometer SD4 (fig. 10 and fig. 11). The landslide displacement rate significantly increases when the groundwater level reaches 8 m in depth (fig. 10 and fig. 11). As soon as this level is reached, several small accelerations are recorded in association to the groundwater fluctuations. Finally, the slope stabilizes when the groundwater drops below 8 m in depth. For the piezometer C2 (fig. 9), close to the GNSS VLRB, the piezometric seasonality is synchronous to the seasonality of the displacements. A piezometric threshold would be 3.10 m depth (fig. 9). For the Geobead probe (C3-A4), set up in July 2010, groundwater levels have been estimated from the pore water pressures recorded by the sensor. The piezometric threshold would be at 3.50 m depth. The differences in depth between the two measurement points can be explained by the difference in altitude (0.21 m) between the two piezometers and by the drawdown of the groundwater approaching the boundaries of the chalk panel.

43As previously mentioned, the definition of thresholds is also based on analysis of hydro-climatic conditions of the four major acceleration events (1982, 1988, 1995 and 2001) to define historical thresholds. As can be observed on figure 9, the major accelerations are associated to a high groundwater level linked to consecutive years of rainfall excess (Bogaard et al., 2011; Lissak, 2012). The critical groundwater level for landslide stability at Danestal piezometer varies between 9.60 m and 10.35 m in depth. For the acceleration of 1988, 1995 and 2001, the groundwater levels are quite similar, between 10.20 and 10.35 m in depth. As a first approximation, we can consider a critical regional groundwater level threshold close to 10.35 m depth. If this threshold is reached or exceeded, a major acceleration can occur. Moreover, we can observe that, regardless of the groundwater level at the velocity peak, the analysis has focused on the groundwater level at the beginning of the hydrological year (in early winter); in these periods, the groundwater level can rise of 1.48 m to 2.10 m in a few months according to the rainfalls and quickly reach a critical level.

44Analysis of the Danestal piezometer allows definition of two significant piezometric thresholds for, respectively, a seasonal acceleration of low amplitude (threshold at -13 m in depth, fig. 8 and fig. 9) and a large acceleration of high amplitude (threshold at 10.35 m in depth, fig. 8). Consequently in-situ observations suggest that a change of +2.65 m of the groundwater level at Danestal is necessary for major acceleration triggering.

Conclusion

45The Cirque des Graves landslide is a typical example of slow-moving and deep-seated landslide developed in a coastal vulnerable area. The hydrological analysis based on historical data and field monitoring highlights the spatial distribution and the temporal variability of the landslide kinematics. Thus, the assessment of the temporal occurrence of the landslide is complex. Statistical methods are generally suggested to investigate the temporal occurrence of landslides. For the moment, only a semi-quantitative approach can be carried out through field surveys. The use of permanent GNSS stations provides infra-centimetric measurements and the quantification of very low displacements of the landslide (movements that are imperceptible by traditional geodetic measurements). However, the installation at several points of sustainable permanent GNSS station with a daily transmission of data is still expensive. It is therefore necessary to implant receivers at strategic points, representative of the whole behaviour of the landslide.

46The combination of repeated campaigns and permanent monitoring is necessary in this case study in order to: 1) highlight the regressive dynamics of the landslide, 2) evaluate the spatial and temporal heterogeneity of the displacement rates, 3) evaluate the effect of groundwater fluctuations on the velocity pattern, 4) define piezometric threshold for landslide reaction. Two significant piezometric thresholds can be defined. A first threshold based on field surveys (short-term/high resolution) for moderate landsliding events recorded between 2009 and 2013 has been defined at 3.50 m depth. This threshold corresponds to a depth of 13 m of the groundwater level on the plateau. A second threshold based on historical data (long-term/low resolution) has been defined at -10.35 m depth for the four major landsliding events that occurred between 1982 and 2001. Thus, by comparing the two thresholds, we can propose a change of groundwater level necessary to trigger a major landsliding event of about 2.65 m if the groundwater level is close to 13 m depth.

47These preliminary results suggest that the effect of rainfall on slope stability is hampered by the prior groundwater levels (several months before a main landsliding event).

Top of page

Bibliography

Aleotti P. (2004) – A warning system for rainfall-induced shallow failures. Engineering Geology 73(3), 247-265.

Ayalew L., Hiromitsu Y., Hideaki M., Takami K. (2005) – Landslides in Sado Island of Japan. Case studies, monitoring techniques and environmental considerations. Engineering Geology 81, 419-431.

Baldi P., Cenni N., Fabris M., Zanutta A. (2008) – Kinematics of a landslide derived from archival photogrammetry and GPS data. Geomorphology 102, 534-444.

Barry R.G. (1981)Mountain, weather and climate. Methuen, London, 313 p.

Bogaard T.A., Antoine P., Desvarreux P., Giraud A., Van Asch T.W.J. (2000) – The slope movements within the Mondorès graben (Drôme, France); the interaction between geology, hydrology and typology. Engineering geology 55, 4, 297-312.

Bogaard T.A., Devi Maharjan L., Maquaire O., Lissak C., Malet J.-P. (2011) – Identification of hydro‐meteorological triggers for Villerville coastal landslide. Proceedings of the Second World Landslide Forum, 3‐7 October 2011, Rome, 5 p.

Brückl E., Brunner F.K., Kraus K. (2006) – Kinematics of a deep-seated landslide derived from photogrammetric gps and geophysical data. Engineering Geology 88, 149-159.

Caine N. (1980) – The rainfall intensity: duration control of shallow landslides and debris flows. Geografiska Annaler, Physical Geography 62, 1, 23-27.

Cappa F., Guglielmi Y., Soukatchoff V.M., Mudry J., Bertrand C., Charmoille A. (2004) – Hydromechanical modeling of a large moving rock slope inferred from slope levelling coupled to spring long-term hydrochemical monitoring: example of the La Clapière landslide (Southern Alps, France). Journal of Hydrology 291, 67-90.

Corominas J., Moya J., Lloret A., Gili J.A., Angeli M.G., Pasuto A., Silvano S. (2000) – Measurement of landslide displacements using a wire extensometer. Engineering Geology 55, 149-166.

Corsini A., Pasuto A., Soldati M., Zannomi A. (2005) – Field monitoring of the Corvara landslide Dolomites Italy and its relevance for hazard assessment. Geomorphology 66, 149-165.

Crozier M.J. (1986)Landslides: Causes, Consequences and Environment, Croom Helm, London and Dover, N.H., 352 p.

Cruden D.M., Varnes D.J. (1996) – Landslide types and processes. In Turner A.K., Schuster R.L. (Eds.) Landslides, Investigation and Mitigation. Special Report 247, Transportation Research Board, Washington, 36-75.

Delmonaco G., Margottini C. (2004) – Meteorological factors influencing slope stability. In Casale R., Margottini C. (Eds.), Natural Disasters and Sustainable Development, Springer Verlag, 19-35.

Demoulin A., Glade T. (2004) – Recent landslide activity in Manaihan Belgium. Landslides, 1, 4, 305-310.

Devi Maharjan L. (2011)A hydrological analysis of the Villerville landslide. Rapport de stage, Unesco-IHE, Institute for Water Education, Delft, the Netherlands, 90 p.

Dixon T.H. (1991) – An introduction to the global positioning system and some geological application. Reviews of Geophysics 29, 2, 249-276.

Flageollet J.C., Helluin E. (1987) – Morphological investigations in the sliding areas along coast of Pays d’Auge, near Villerville, Normandy, France. In Gardiner V. (Ed.) International Geomorphology 1986, Proceedings of the First International Conference on Geomorphology, Wiley J. and Sons Ltd., Chichester, New York, 447-486.

Floris M., Bozzano F. (2008) – Evaluation of landslide reactivation: A modified rainfall threshold model based on historical records of rainfall and landslides. Geomorphology 94, 40-57.

Gili J.A., Corominas J., Rius J. (2000) – Using Global Positioning System techniques in landslide monitoring. Engineering Geology 55, 167-192.

Gonzalez-Garcia A.J., Mayorga-Marquez R. (2004) – Thresholds for rainfall events that induce landslides in Colombia. In Lacerda W., Ehrlich M., Fontoura S.A.B., Sayao A.S.F. (Eds.) Landslides: Evaluation and Stabilization, Taylor and Francis Group, London, 349-355.

Guzzetti F., Cardinali M., Reichenbach P., Cipolla F., Sebastiani C., Galli M., Salvati P. (2004) – Landslides triggered by the 23 November 2000 rainfall event in the Imperai Province, Western Liguria, Italy. Engineering Geology 73, 1-17.

Guzzetti F., Peruccacci S., Rossi M., Stark C.P. (2008) – The rainfall intensity-duration control of shallow landslides and debris flows: an update. Landslide 5, 3-17.

Hansen A. (1984) – Landslide hazard analysis. In Brunsden D., Prior D.B. (Eds.) Slope Instability. Wiley, New York, 523-602.

Iverson R.M. (2000) – Landslides triggering by rain infiltration. Water Resources Research 36, 7, 1897-1910.

Jackson M.E., Bodin P.W., Savage W.Z., Nel E.M. (1996)Measurement of local horizontal velocities on the Slumgullion landslide using the global positioning system. Open File Report 1996-2130, U.S. Geological Survey, 93-95.

Lissak C. (2012)Les glissements de terrain côtiers du Pays d’Auge (Calvados) : Morphologie, fonctionnement et gestion du risque. Thèse de doctorat, Université de Caen Basse-Normandie, 312 p.

Lissak C., Maquaire O., Malet J.-P. (2009) – Role of hydrological process in landslide occurrence: Villerville-Cricqueboeuf landslides Normandy coast, France. In Malet J.-P., Remaître A., Boogard T.A. (Eds.) Proceedings of the International Conference on Landslide Processes: from Geomorphologic Mapping to Dynamic Modeling. Strasbourg, CERG, 175-180.

Lissak C., Maquaire O., Malet J.-P. (2010) – Multi-technique permanent monitoring of a slow-moving coastal landslide in Normandy. In Malet J.-P., Remaître A., Boogard T.A. (Eds.) Proceedings of the International Conference: ' Mountain Risks: Bringing Science to Society '. Strasbourg, CERG, 267-274.

Lissak C., Puissant A., Maquaire O., Malet J.-P. (2013) – Analyse spatio-temporelle de glissements de terrain littoraux par l’exploitation de données géospatiales multi-sources. Revue Internationale de Géomatique, 23, 2, 199-225.

Lissak C., Maquaire O., Malet J.-P., Bitri A., Samyn K., Grandjean G., Bourdeau C., Reiffsteck P., Davidson R. (2014) – Airborne and ground-based sources of information for characterizing the morphostructure of coastal landslides. Geomorphology, in press.

Malet J.-P., Hartig S., Calais E., Maquaire O. (2000) – Apport du GPS au suivi en continu des mouvements de terrain. Application au glissement-coulée de Super-Sauze, Alpes-de-Haute-Provence, France. Comptes-Rendus de l’Académie des Sciences, Sciences de la Terre et des Planètes, 331, 175-182.

Malet J.-P., Maquaire O., Calais E. (2002a) – The use of Global Positioning System techniques for the continuous monitoring of landslide: application to the Super-Sauze earthflow, Alpes-de-Haute-Provence, France, Geomorphology 43, 33-54.

Malet J.-P., Maquaire O., Calais E. (2002b) – Le GPS en géomorphologie dynamique. Application à la surveillance de mouvements de terrain Super-Sauze, Alpes du Sud, France. Géomorphologie : relief, processus, environnement, 2, 165-180.

Malet J.-P., van Ash T.W.J., van Beek L.P.H., Maquaire O. (2003a) – Apport des modèles hydrologiques spatialisés à la simulation numérique de glissements de terrain. Impacts pour la gestion du risque. Journée pour la Prévention des Risques, BRGM, Orléans, Janvier 2003, 7 p.

Malet J.-P., Maquaire O., van Ash T.W.J. (2003b) – Hydrological behaviour of earthflows developed in clay-shales: investigation, concept and modelling. Proceeding of the International Workshop on Occurrence and Mechanisms of Flow-like Landslides in Natural Slopes and Earthfills, Sorrento, Italy, 14-16.

Malet J.-P., Ulrich P., Déprez A., Masson F., Lissak C., Maquaire O., (2013) – Continuous monitoring and near-real time processing of GPS observations for landslide analysis: a methodological framework. In Margottini, C., Canuti, P., Sassa, K. (Eds.), Landslide Science and Practice, Volume 2: Early Warning, Instrumentation and Monitoring, 201-209.

Maquaire O., (1990) Les mouvements de terrain de la côte du Calvados. Recherches et prévention. Documents du BRGM 197, BRGM, Orléans, 430 p.

Maquaire O. (2000) – Effects of groundwater on the Villerville-Cricqueboeuf landslide. Sixteen years of survey, Calvados, France. Proceedings of the 8th International Symposium on Landslides, Cardiff, Telford, 2, 1005-1010.

Mora P., Baldi P., Casula G., Fabris M., Ghirotti M., Mazzini E., Pesci A. (2003) – Global Positioning Systems and digital photogrammetry for the monitoring of mass movements: application to the Ca' di Malta landslide northern Apennines, Italy. Engineering Geology 68, 103-121.

Moss J.L. (2000) – Using Global Positioning System to monitor dynamic ground deformation networks on potentially active landslides. International Journal of Applied Earth Observation and Geoinformation 2-1, 24-32.

Peters E.T., Malet J.-P., Bogaard T.A. (2010) – Multi-sensor monitoring network for real-time landslide forecasts in early warning systems. In Malet J.-P., Remaître A., Boogard T.A. (Eds.) Proceedings of the International Conference: 'Mountain Risks: Bringing Science to Society'. Strasbourg, CERG, 335-340.

Petley D.N., Mantovani F., Bulmer M.H., Zannoni A. (2005) – The use of surface monitoring data for the interpretation of landslide movement patterns. Geomorphology 66, 133-147.

Peyret M., Djamour Y., Rizza M., Ritz J.F., Hurtrez J.E., Goudarzi M.A., Nankali H., Chéry J., Le Dortz K., Uri F. (2008) – Monitoring of the large slow Kahrod landslide in Alborz mountain range Iran by GPS and sar interferometry. Engineering Geology 100, 131-141.

Soeters R., van Westen C.J. (1996) – Slope instability recognition, analysis, and zonation. In Turner K.A., Schuster R.L. (Eds.) Landslides: investigation and mitigation. Transport Research Board Special Report 247, 129-177.

Squarzoni C., Delacourt C., Allemand P. (2005) – Differential single-frequency GPS monitoring of the La Valette land-slide, French Alps. Engineering Geology 79, 215-229.

van Asch T.W.J., Buma J. (1996) – Modelling groundwater fluctuations and the frequency movement of a landlside in the Terres Noires region of Barcelonette, France. Earth surface processes and landforms 22, 131-141.

van Asch T.W.J, Buma J., van Beek L.P.H. (1999) – A view on some hydrological triggering systems in landslides. Geomorphology 30, 25-32.

Zhu X., Xu Q., Zhou J., Deng M. (2012) – Remote Landslide Observation System with Differential GPS. Procedia Earth and Planetary Science 5, 70-75.

Top of page

Annex

Version française abrégée

La surveillance des glissements de terrain est fondée sur l’observation et l’analyse croisée des différents paramètres qui régissent l’instabilité du versant (la piézométrie, la météorologie…). Traditionnellement, la prédiction des mouvements de terrain est fondée sur l’analyse de données historiques (Malet et al., 2003a ; Guzzetti et al., 2004) et sur des investigations de terrain (Hansen, 1984 ; Soeters et van Westen, 1996 ; Malet et al., 2003b) permettant d’observer à la fois les forçages hydro-climatiques et la cinématique de glissement de façon à déterminer des seuils hydro-climatiques dans le déclenchement des instabilités (Caine, 1980 ; Corominas et al., 2000 ; Iverson, 2000 ; Aleotti, 2004 ; Guzzetti et al., 2004 ; Guzzetti et al., 2008).

Afin de préciser leur rôle respectif, chacun des paramètres peut être observé, mesuré en continu ou lors de campagnes répétées de mesures. Le choix des techniques employées et la configuration du réseau de surveillance dépend du type de glissement étudié et de la précision recherchée (Moss, 2000). De nombreuses techniques de positionnement conventionnelles appliquées aux glissements de terrain sont disponibles (Corominas et al., 2000 ; Malet et al., 2002a ; Mora et al., 2003 ; Ayalew et al., 2005 ; Petley et al., 2005 ; Squarzoni et al., 2005). Ces différentes techniques géodésiques, qui peuvent être complémentaires, permettent une approche de la cinématique de versant à différentes échelles spatio-temporelles et offrent différentes précisions de mesures. Les outils géodésiques classiques sont depuis une quinzaine d'années associés ou remplacés par l'utilisation de techniques GNSS qui permet de positionner de façon relative et en temps réel les coordonnées de cibles (Jackson et al., 1996 ; Corominas et al., 2000 ; Gili et al., 2000 ; Moss, 2000 ; Malet et al., 2002a et 2002b ; Mora et al., 2003 ; Demoulin et Glade, 2004 ; Corsini et al., 2005 ; Peyret et al., 2008).

Plusieurs études ont démontré, comme dans le cas présent, que l’amplitude des déplacements était directement contrôlée par les apports pluviométriques (van Asch et al., 1999 ; Delmonaco et Margottini, 2004) et par les fluctuations de la nappe souterraine (Bogaard et al., 2000 ; Cappa et al., 2004).

Les études précédentes menées sur le glissement du Cirque des Graves le long des côtes de Basse-Normandie, ont montré que la dynamique du glissement était largement conditionnée par les forçages hydro-climatiques. Dans une approche prospective, l’un des objectifs était alors de déterminer des niveaux pluviométriques et piézométriques critiques à la mise en mouvement saisonnière du glissement ou lors d'accélérations majeures. Ainsi nous nous sommes focalisés sur les conditions hydro-météorologiques avant, pendant et après les accélérations majeures du Cirque des Graves (janvier 1982, février 1988, mars 1995, mars 2001) qui ont provoqué des décrochements de plusieurs décimètres et de nombreux dégâts.

Avec le réseau de surveillance mis en place (Lissak et al., 2010), nous disposons d’une base de données multi-paramètres et multi-temporelles qui permet une analyse croisée de la cinématique, de la pluviométrie et des précipitations efficaces permettant de comprendre le comportement du glissement.

Un réseau de repères topographiques utilisé pour mesurer la cinématique du glissement a été implanté de façon à répartir les points de mesures sur différents compartiments qui composent le glissement (Lissak et al., 2013). Le réseau s'étend sur environ 1 km d’est en ouest et est composé de vingt-trois bornes-repères. Depuis 2008, plusieurs campagnes de mesures ont été réalisées par GPS. Le temps de session pour chacun des repères a été évalué à 17 minutes pour obtenir un résultat infra-centimétrique. Pour compléter le réseau et obtenir des informations sur l'évolution du glissement en temps réel avec des erreurs de positionnement moindres, trois récepteurs GNSS (Global Navigation Satellite System) permanents ont été installés en juillet 2009. La position de chaque récepteur est enregistrée toutes les trente secondes pour des sessions de vingt-quatre heures et fournit des solutions automatiques en quasi temps réel (j+ 2 jours) disponibles en ligne (Malet et al., 2013).

La définition de seuils de déclenchement des accélérations nécessite de longues séries de données piézométriques au moins antérieures à la première crise de 1982 sur des piézomètres toujours surveillés. Aucun piézomètre n’étant installé directement dans la zone instable avant 1985, un piézomètre issu du réseau Bassin Seine-Normandie du BRGM, a été sélectionné sur le plateau en amont de la zone car il offre des données pertinentes sur la nappe de la craie depuis 1974. Par la suite, plusieurs forages ont été réalisés dans la zone en glissement, où la nappe est fragmentée (Lissak, 2012), afin d'obtenir des mesures ponctuelles ou continues de niveaux de nappe et de pressions interstitielles in-situ. Ainsi, au total, le réseau de suivi piézométrique est composé de 29 points d’observation dont 6 en continu. Le réseau de surveillance climatique est constitué de deux stations dont une station Météo-France implantée depuis 1949 sur le plateau, en amont du glissement, et une autre implantée à proximité du glissement depuis 2013.

Depuis 2008, les mesures sont menées en période d’activité réduite, sans accélération majeure du glissement. Les résultats des observations confirment l’activité permanente du glissement du Cirque des Graves avec des déplacements saisonniers de quelques centimètres par an. Les récepteurs GNSS ont largement permis de préciser des valeurs de déplacement infra-centimétriques et donc de détecter les mouvements imperceptibles à l’œil, de mettre en évidence un décalage temporel de plusieurs jours dans la mise en mouvement des compartiments sur lesquels sont implantés les récepteurs et de confirmer le déclenchement des instabilités par l’aval. Cette approche méthodologique quantitative est adéquate pour les instabilités dont l’activité est très réduite comme au Cirque des Graves et pour analyser la cinématique du glissement en réponse aux facteurs de déclenchement dans le cadre de la mise en place de système d'alerte. Cependant, la conjugaison de campagnes ponctuelles avec les mesures de déplacements en continu est nécessaire pour mettre en avant l’hétérogénéité spatiale des champs de déplacements et pour souligner le compartimentage du glissement en diverses unités morphodynamiques.

Une première analyse a permis d'estimer de façon empirique des seuils piézométriques régionaux historiques pour les accélérations majeures de 1982, 1988, 1995 et 2001. Ces accélérations se sont déclenchées au cours d'une situation hydro-climatique spécifique marquée par des niveaux de nappe phréatique très élevés en relation directe avec une pluviométrie efficace pluriannuelle bien supérieure à la moyenne (Bogaard et al., 2011 ; Lissak, 2012). En première approximation, il est possible de retenir une profondeur de nappe critique à -10,35 m sur le plateau en amont (seuil régional). Si ce niveau est atteint, des accélérations brutales peuvent se déclencher.

La précision des relations nappe-déplacements et la définition de seuils locaux est fondée sur l’analyse des mesures de terrain à haute résolution temporelle et spatiale. Depuis la mise en place des récepteurs GPS en 2009, quatre périodes d'accélération mineures ont été mises en évidence avec des déplacements pouvant atteindre plus de 8 cm. La saisonnalité des déplacements a été mise en relation étroite avec les périodes de recharge de la nappe, avec un décalage dans le temps entre la montée des eaux souterraines et l'accélération du glissement entre 1 et 4 jours.

Top of page

List of illustrations

Title Fig. 1 – Location and geomorphological setting of the Cirque des Graves landslide. Fig. 1 – Localisation et contexte géomorphologique du glissement du Cirque des Graves.
Caption A: topographic context with location of the active landslide along the coast between Trouville and Honfleur. B: 3D view of the Cirque des Graves landslide in 2006. C: geological cross-section. 1: altitude; 2: coastal slope; 3: Pays d’Auge plateau; 4: active coastal landslide; 5: road D513; 6: buildings; 7: surficial deposit; 8: chalk; 9: glauconitic sand; 10: marls; 11: sandstone; 12: spring; 13: sliding surface.A : contexte topographique régional et localisation des glissements de terrain actifs le long de la côte entre Trouville et Honfleur. B : vue 3D du glissement du Cirque des Graves en 2006. C : coupe géologique schématique. 1 : point côté ; 2 : versant côtier ; 3 : plateau du Pays d’Auge ; 4 : glissement de terrain actif ; 5 : route départementale 513 ; 6 : bâti ; 7 : formations superficielles ; 8 : craie ; 9 : sable glauconieux ; 10 : marne ; 11 : grès ; 12 : résurgence ; 13 : surface de glissement.
URL http://journals.openedition.org/geomorphologie/docannexe/image/10607/img-1.png
File image/png, 1.8M
Title Fig. 2 – Monitoring network for the Cirque des Graves landslide. Fig. 2 – Réseau de surveillance géodésique du Cirque des Graves.
Caption A: Cirque des Graves landslide. B: Pays d’Auge plateau. 1: piezometer; 2: wells; 3: inclinometer; 4: piezometer with permanent sensor; 5: geobead probe; 6: meteorological station; 7: topographic benchmark; 8: permanent GNSS receivers; 9: active landslide; 10: Cirque des Graves landslide.A : Cirque des Graves. B : plateau du Pays d’Auge. 1 : piézomètre ; 2 : puits ; 3 : inclinomètre ; 4 : piézomètre avec capteur permanent ; 5 : sondes Geobead ; 6 : station météorologique ; 7 : repère topographique ; 8 : récepteurs GNSS permanents ; 9 : glissement de terrain actif ; 10 : glissement du Cirque des Graves.
URL http://journals.openedition.org/geomorphologie/docannexe/image/10607/img-2.png
File image/png, 1.7M
Title Fig. 3 – Annual average displacement rate at Cirque des Graves landslide between 2 April 2008 and 25 April 2012. Fig. 3 – Vitesses moyennes annuelles des déplacements mesurés au Cirque des Graves entre le 2 avril 2008 et le 25 avril 2012.
Caption 1: average annual velocity; 2: topographic benchmark; 3: permanent GNSS receiver; 4: limit of the active zone; 5: coast line; 6: chalk panel limit; 7: location of the schematic profile of fig. 4. 1 : vitesse moyenne annuelle ; 2 : repère topographique ; 3 : récepteur GNSS permanent ; 4 : limite de la zone active ; 5 : trait de côte ; 6 : limite de panneau de craie ; 7 : localisation du profil schématique de la fig. 4.
URL http://journals.openedition.org/geomorphologie/docannexe/image/10607/img-3.png
File image/png, 895k
Title Fig. 5 – Cumulative displacements since 1985 of benchmarks located on the eastern part of the Cirque des Graves landslide. Fig. 5 – Déplacements cumulés mesurés depuis 1985 des repères situés dans le secteur est du Cirque des Graves.
URL http://journals.openedition.org/geomorphologie/docannexe/image/10607/img-4.png
File image/png, 516k
Title Fig. 6 – VLRH-VLRV and VLRB-VLRV baselines evolution between August 2009 and April 2013.Fig. 6 – Longueurs des lignes de base VLRH-VLRV et VLRB-VLRV enregistrées entre août 2009 et avril 2013.
URL http://journals.openedition.org/geomorphologie/docannexe/image/10607/img-5.png
File image/png, 314k
Title Fig. 7 – Rainfall – groundwater levels (at SD4) relationships during a typical recharge period. Fig. 7 – Relation pluie-nappe (à SD4) en période de recharge.
URL http://journals.openedition.org/geomorphologie/docannexe/image/10607/img-6.png
File image/png, 404k
Title Fig. 8 – Relation between groundwater level variations at Danestal piezometer and effective rainfall between at Saint-Gatien between 1974 and 2013. Fig. 8 – Relations entre les variations de la nappe de Danestal et les pluies efficaces annuelles entre 1974 et 2013.
Caption A: groundwater level at Danestal piezometer since 1974. B: excessive rainfall compared to the mean annual effective rainfall at Saint-Gatien station. A : fluctuation de la nappe enregistrée au piézomètre Danestal depuis 1974. B : excédent pluviométrique par rapport à la moyenne annuelle des précipitations efficaces enregistrées à la station de Saint-Gatien.
URL http://journals.openedition.org/geomorphologie/docannexe/image/10607/img-7.png
File image/png, 556k
Title Fig. 9 – Relations between the displacements (GNSS VLRH and VLRB), piezometric levels (SD4 and C2), pore pressure (A4) recorded at Cirque des Graves and daily rainfall of Saint-Gatien station. Fig. 9 – Relations entre les déplacements GNSS (VLRH et VLRB), les niveaux piézométriques (SD4 et C2), les pressions interstitielles (A4) enregistrés au Cirque des Graves et les précipitations journalières de la station de Saint-Gatien.
URL http://journals.openedition.org/geomorphologie/docannexe/image/10607/img-8.png
File image/png, 703k
Title Fig. 10 – Relations between piezometric fluctuations (SD4), surface displacements (VLRH and VLRB) and cumulative effective rainfall during the first period of acceleration (20 January 2010 to 13 May 2010).Fig. 10 – Relations entre les fluctuations piézométriques (SD4), les déplacements de surface (VLRH et VLRB) et la pluie efficace cumulée durant la 1ère période d'accélération (20 janvier 2010 au 13 mai 2010).
URL http://journals.openedition.org/geomorphologie/docannexe/image/10607/img-9.png
File image/png, 389k
Title Fig. 11 – Relations between piezometric fluctuations (SD4), surface displacements (VLRH and VLRB) and cumulative effective rainfall during the fourth period of acceleration started in early December 2012.Fig. 11 – Relations entre les fluctuations piézométriques (SD4), les déplacements de surface (VLRH et VLRB) et la pluie efficace cumulée durant la quatrième période d'accélération démarrée en début décembre 2012.
URL http://journals.openedition.org/geomorphologie/docannexe/image/10607/img-10.png
File image/png, 1.1M
Top of page

References

Bibliographical reference

Candide Lissak, Olivier Maquaire, Robert Davidson and Jean-Philippe Malet, Piezometric thresholds for triggering landslides along the Normandy coast, FranceGéomorphologie : relief, processus, environnement, vol. 20 - n° 2 | 2014, 145-158.

Electronic reference

Candide Lissak, Olivier Maquaire, Robert Davidson and Jean-Philippe Malet, Piezometric thresholds for triggering landslides along the Normandy coast, FranceGéomorphologie : relief, processus, environnement [Online], vol. 20 - n° 2 | 2014, Online since 01 January 2016, connection on 18 April 2024. URL: http://journals.openedition.org/geomorphologie/10607; DOI: https://doi.org/10.4000/geomorphologie.10607

Top of page

About the authors

Candide Lissak

Laboratoire LETG-Caen GEOPHEN – Géographie Physique et Environnement – UMR 6554 CNRS – Université de Caen-Basse-Normandie – Esplanade de la Paix – 14032 Caen cedex – France (candide.lissak-borges@unicaen.fr).

By this author

Olivier Maquaire

Laboratoire LETG-Caen GEOPHEN – Géographie Physique et Environnement – UMR 6554 CNRS – Université de Caen-Basse-Normandie – Esplanade de la Paix – 14032 Caen cedex – France.

By this author

Robert Davidson

Laboratoire LETG-Caen GEOPHEN – Géographie Physique et Environnement – UMR 6554 CNRS – Université de Caen-Basse-Normandie – Esplanade de la Paix – 14032 Caen cedex – France.

By this author

Jean-Philippe Malet

Institut de Physique du Globe de Strasbourg – CNRS UMR 7516 – Ecole et Observatoire des Sciences de la Terre – EOST/Université de Strasbourg – 5 rue Descartes – F-67084 Strasbourg – France.

By this author

Top of page

Copyright

The text and other elements (illustrations, imported files) are “All rights reserved”, unless otherwise stated.

Top of page
Search OpenEdition Search

You will be redirected to OpenEdition Search