Search

International audience; Between October 2010 and November 2017, surveys using terrestrial laser scanning (TLS) were carried out every 4-5 months to monitor erosion on three chalk cliff faces with close structural characteristics in Normandy (the two Dieppe sites are abandoned cliffs while the Varengeville-sur-Mer site is an active one). The uniqueness of this paper is the comparison between these cliff sections and the significance of marine and/or subaerial controls in erosion. Beyond the quantification of annual retreat rates, which on the active cliff are currently 36 times greater than on the abandoned cliffs, this paper highlights the fundamental efficacy of marine controls in the current cliff erosion and the removal of falls (analyses of modalities of retreat and profile evolution according to a naturalistic approach). Subaerial processes appear to have a less erosive action than marine ones, because the studied abandoned cliffs show low erosion dynamics and have maintained a subvertical profile despite continentalization for between 30 to more than 120 years.

Images from agile (viewing angle over 40 degrees) and very high spatial resolution satellites (inferior to 1 m) can be useful for monitoring cliff faces, which is the best proxy to better understand coastal cliff dynamics. However, these images with a specific configuration are rarely used, partly because it is cumbersome to process them. Based on Pleiades images of the coastal cliff face along the coast of Normandy, with a high angle of incidence (up to 40 degrees) and taken on multiple dates, the paper aims to identify i) the best open-source processing chain to reconstitute three-dimensional (3D) cliff faces by stereo restitution ii) the reasons behind its best performance and iii) the key parameters to change depending on the image datasets or processing chains so as to facilitate transposition. The Ames Stereo Pipeline (R) (ASP (R)) and MicMac (R) software programmes were tested using different parameters (matching algorithm, size of correlation window, etc.) for the 3D reconstructions. MicMac (R) provides the best performance using GeomImage (1-2 pixel matching) with a size of correlation window of 3 x 3 or 7 x 7 associated with a regularization parameter of 0.10. With these parameters, the point clouds of the cliff face have an average point density of 1.70 point m(-2), a mean distance from Unmanned Aerial Vehicle (UAV) ground truth data of 0.04 m and a standard deviation of 1.72 m. With these characteristics, the threshold of rockfall detection using a multi-source comparison is assessed at 100 m(3), which involves that the large majority of rockfalls (69%) around the study area could be detected by a diachronic approach. Considering the daily Pleiades revisiting time, this method offers a great opportunity to monitor erosion and to better understand coastal cliff dynamics.

International audience; 21 To monitor chalk cliff face along the Normandy coast (NW France) which is prone to erosion, 22 we tested the potential of cliff face 3D reconstruction using pairs of images with high angle of 23 incidence at different dates from the agile Pléiades satellites. The verticality aspect of the cliff 24 face brings difficulties in the 3D reconstruction process. Furthermore, the studied area is 25 challenging mainly because the cliff face is north-oriented (shadow). Pléiades images were 26 acquired over several days (multi-date stereoscopic method) with requested incidence angles 27 until 40°. 3D reconstructions of the cliff face were compared using two software: ASP ® and 28 ERDAS IMAGINE ®. Our results are twofold. Firstly, despite ASP ® provides denser point clouds 29 than ERDAS IMAGINE ® (an average of 1.60 points/m² from 40° incidence angle stereoscopic 30 pairs on the whole cliff face of Varengeville-sur-Mer against 0.77 points/m² respectively), 31 ERDAS IMAGINE ® provides more reliable point clouds than ASP ® (precision assessment on 32 the Varengeville-sur-Mer cliff face of 0.31 m ± 2.53 and 0.39 m ± 4.24 respectively), with a 33 better spatial distribution over the cliff face and a better representation of the cliff face shape. 34 Secondly, the quality of 3D reconstructions depends mostly on the amount of noise from raw 35 images and on the shadow intensity on the cliff face (radiometric quality of images). 36 37 Keywords 38 Pléiades imagery; high incidence angle images; cliff face erosion; 3D reconstruction; 39 Normandy 40 41

UAVs are relevant for monitoring cliff faces. In this study, several flights are performed with various imaging angles (nadir, 20 degrees, 30 degrees and 40 degrees off-nadir) to assess the impact of the imaging angle on the 3D cliff face reconstruction. Occlusions issues arising with sub-vertical cliffs make nadir surveys nearly irrelevant. The results obtained with 20 degrees, 30 degrees and 40 degrees off-nadir imaging angles are satisfactory regarding texture restitution and accuracy with respectively 5.5 cm, 5.9 cm and 4.9 cm of error, higher tilting angles yielding better reconstructions on sub-vertical or overhanging parts of the cliff. This article also investigates other parameters affecting tiepoint detection on the cliff face, as the effective overlap, the UAV-cliff face distance and the cliff face illumination. Guidelines are provided for UAV survey parameterization, aiming at capturing the whole cliff face with a good trade-off between distance to the cliff, flight height and spatial resolution of the photographs.

High-resolution digital elevation models (HRDEMs) on coastal areas, particularly cliffs, are an important asset for analyzing and quantifying the processes affecting their morphology. The Vaches-Noires coastal cliffs, located at the north-western part of the Pays-d’Auge region (Normandy, France), are an ideal site for the study of subaerial and marine erosion processes. Indeed, the coastline evolves through the accumulation of deposits resulting from rotational landslides and/or mudflows at the base of the cliff which are then undermined by the sea. The 3D modeling of a representative sector of the cliffs appears a suitable solution to monitor and quantify this evolution using HRDEMs. This paper presents the whole of a monitoring protocol, with high spatial resolution and high temporal resolution, on an original location characterized by reliefs of Badlands in coastal environments. The first results of these investigations are described with an evaluation of two methods for the creation of 3D models: a “heavy” method by Terrestrial Laser Scanning (TLS) monitoring, and a “light” method using “Structure from Motion” (SfM) terrestrial photogrammetry. The small investment of time in the field and money is an important asset of the SfM method. The textured rendering quality of the SfM model makes it a powerful tool for geomorphological analysis. According to a recurrent monitoring protocol involving both measurements, it was possible to compare the accuracy of the SfM and TLS models. The results show a high accuracy of the SfM models with an overall average error of 0.05 m. The representations of the zones of accumulation and erosion as well as their spatial succession allow to investigate the dynamics of the involved processes with both methods. Preliminary volume calculation results are conclusive for TLS models.

International audience; 18 This paper aims to compare models from terrestrial laser scanning (TLS), 19 terrestrial photogrammetry (TP), and unmanned aerial vehicle photogrammetry 20 (UAVP) surveys to evaluate their potential in cliff erosion monitoring. TLS has 21 commonly been used to monitor cliff-face erosion (monitoring since 2010 in 22 Normandy) because it guarantees results of high precision. Due to some 23 uncertainties and limitations of TLS, TP and UAVP can be seen as alternative 24 methods. First, the texture quality of the photogrammetry models is better than 25 that of TLS which could be useful for analysis and interpretation. Second, a 26 comparison between the TLS model and UAV or TP models shows that the mean 27 error value is mainly from 0.013 to 0.03 m, which meets the precision 28 requirements for monitoring cliff erosion by rock falls and debris falls. However, 29 TP is more sensitive to roughness than UAVP, which increases the data standard 30 deviation. Thus, UAVP appears to be more reliable in our study and provides a 31 larger spatial coverage, enabling a larger cliff-face section to be monitored with a 32 regular resolution. Nevertheless, the method remains dependent on the weather 33 conditions and the number of operators is not reduced. Third, even though UAVP 34 has more advantages than TP, the methods could be interchangeable when no 35 pilot is available, when weather conditions are bad or when high reactivity is 36 needed. 37

The triggering factors of rock falls remain unknown due to a lack of exhaustive, regular and accurate surveys. Based on an inventory of 331 rock falls collected weekly between 2002 and 2009 from Veules-les-Roses to Le Treport (Upper Normandy), the relationships between coastal chalk cliff rock falls (dates and geomorphological features) and external factors commonly agreed as triggering (rainfall, temperature variations, tide and wind) are studied. The combination of multivariate statistical and empirical analyses indicates that (1) ‘cold and dry weather’ and ‘high rainfall and high wind’ are the conditions most likely to trigger rock falls, (2) the main triggering factors of rock falls are effective rainfall (for rock falls mostly between 200 and 1400 m3 or larger than 10 000 m3 and coming from the whole cliff face), freeze/thaw cycles (especially for rock falls smaller than 200 m3 and coming from the foot and top of the cliff face) and marine roughness (rock falls mainly smaller than 200 m3 and coming from the cliff foot). However, the contribution of each factor to triggering is difficult to determine because of combinations of factors (85% of 331 cases), relays of processes and hysteresis phenomena. In view of these first results, it is still presumptuous to predict the location and time of triggering of rock falls. However, the statistical and naturalistic approaches adopted and the observations made in this study are from an original database, and constitute a real starting point for the prediction and prevention of the hazard of coastal chalk cliff rock falls in Upper Normandy. Copyright © 2015 John Wiley & Sons, Ltd.

International audience; The georeferencing process is crucial to the accuracy of Terrestrial Laser Scanner data, in particular in the context of diachronic studies relying on multi-temporal surveys. The use of Ground Control Points in the georeferencing process can however be complex when confronted with the practical constraints of coastal surveying.A simple and quick alternative method called “pseudo-direct georeferencing” is proposed in the present paper. This method involves internal inclinometers to measure roll and pitch angles and a centimetric GPS to measure the position of the TLS center and the position of one backsight target. When assessing the transformational uncertainty by using a set of independent ground validation points for both classical indirect and proposed pseudo-direct methods, we respectively obtain Root Mean Square errors of 4.4 cm for the indirect method and 3.8 cm for the pseudo-direct method

International audience; Coastal chalk cliff falls in Upper Normandy (France) are frequent and have specific spatial and temporal distributions. From 2002 to 2009, the ESTRAN organization (Scientific and Technical Space of Aquatic Resources and Navigation) conducted a weekly inventory of 331 cliff falls (location, volume, and date) between Veules-les-Roses and Le Treport (37.5 km). An amount of 331 falls was counted (date, location, measures of fall deposit). This database is remarkable because of the duration and high frequency of field surveys. The geographical and spatial statistics (locational analysis) methods used in this study aim to determine and understand the spatial and temporal distributions of coastal chalk cliff falls. Exhaustive cartography of such falls thus highlights the cap d'Ailly sector as being the most sensitive to an erosive dynamic along the coast from Veules-les-Roses to Le Treport. Furthermore, this cartography stresses many types of fall kinematics. This can be partly explained by lithological characteristics may predispose mass and particle movements. Using the combination of centrographic statistics, Ripley's K-function, Besag's L-function and kernel density, we have identified: (1) high cliff fall polarization at cap d'Ailly; (2) organization scales of fall locations per water year; (3) areas of massive and numerous falls (fall hazard areas). This information is essential for a better understanding of the working of the studied area and also for the set up of risk prevention tools.; Les éboulements/écroulements de falaises en Haute-Normandie sont fréquents et connaissent une répartition spatiale et temporelle spécifique. De 2002 à 2009, l'association ESTRAN (Espace Scientifique et Technique des Ressources Aquatiques et de la Navigation) a conduit un inventaire hebdomadaire des éboulements/écroulements entre Veules-les-Roses et Le Tréport (37,5 km). Au total, 331 événements datés, localisés et cubés furent recensés. Cette base de données est remarquable par la durée et la répétitivité des levés. L'approche géographique et de statistiques spatiales adoptées (analyse spatiale) visent à déterminer et à comprendre la répartition spatiale et temporelle des éboulements/écroulements de falaises crayeuses. Ainsi, la cartographie exhaustive des événements met en évidence qu'entre Veules-les-Roses et Le Tréport, ces derniers se concentrent au cap d'Ailly. Par ailleurs, cette cartographie montre l'existence d'une grande variété de cinématiques d'éboulements/écroulements. Cela s'explique notamment par des caractéristiques lithologiques favorables au déclenchement des mouvements particulaires et de masse. Grâce à la combinaison des statistiques centrographiques, des fonctions K-Ripley et L-Besag, ainsi que du lissage par la méthode des noyaux, nous avons pu identifier : (1) la forte polarisation des éboulements/écroulements au cap d'Ailly, (2) les échelles d'organisation des localisations d'éboulements/écroulements par année hydrique ; (3) les zones à éboulements/écroulements massifs et nombreux. Ces informations sont fondamentales pour une meilleure compréhension du fonctionnement du milieu étudié mais également pour la mise en place d'outils de prévention du risque.

Les falaises de Haute-Normandie connaissent une intense dynamique régressive qui a surtout été étudiée à travers le suivi du haut de falaise. La comparaison de l’orthophotographie de 2008 (BD Ortho®, IGN) avec l’analyse photogrammétrique initiale (1966-1995 ; Costa, 2000) fournit des résultats sur une période plus longue (1966-2008, soit 42 ans). Ces vitesses ont été calculées tous les 50 m pour mettre en évidence des phénomènes de grande échelle, notamment l’influence des obstacles comme les éboulements/écroulements majeurs ou les jetées portuaires, induisant une multiplication par trois à quatre de l’ablation. Au-delà de l’approche classique de l’estimation des vitesses de retrait, une évaluation des modalités de recul des falaises crayeuses (fréquence et intensité) a été menée entre Étretat et Le Tréport grâce au suivi des éboulements/écroulements recensés sur les photographies aériennes verticales de l’IGN préalablement géoréférencées (1939-1995) et sur les orthophotographies (2000 et 2008). Un inventaire des éboulements/écroulements entre 2002 et 2009 (association ESTRAN) et sur 37,5 km de côte complète les informations obtenues. Associées aux vitesses d’ablation, ces valeurs permettent de préciser les modalités d’évolution de ces côtes à falaises. Schématiquement, deux secteurs aux dynamiques distinctes sont observés. Le premier secteur (cap d’Antifer/Saint-Valéry-en-Caux et Dieppe/Le Tréport) a des vitesses de recul faibles à modérées (de 0,09 à 0,18 m/an) et se caractérise par des éboulements/écroulements rares mais volumineux (médiane de l’aire de départ de l’éboulement/écroulement de 700 m² pour Fécamp/Életot). Le second secteur (Saint-Valéry-en-Caux/Dieppe) est affecté par des vitesses de recul plus importantes (0,23 m/an), et les éboulements/écroulements y sont plus fréquents mais peu volumineux (médiane de l’aire de départ de l’éboulement/écroulement de 200 m² pour Quiberville/Dieppe). Grâce au Service Littoral de l’association ESTRAN, l’inventaire hebdomadaire des éboulements/écroulements réalisé entre Veules-les-Roses et Le Tréport (2002-2009) fournit statistiquement des éléments de réponse quant à la fréquence d’occurrence et les périodes de retour des éboulements/écroulements (ajustements à la loi de puissance). Ces valeurs corroborent les mesures de recul établies par photo-interprétation puisque là encore, le secteur du cap d’Ailly apparaît comme le plus dynamique : un éboulement/écroulement de plus de 1 000 m3 par km linéaire est attendu tous les 6 mois contre tous les 2 à 3 ans pour les autres secteurs (Veules-les-Roses/Le Tréport, Hautot-sur-Mer/Bracquemont). Les informations obtenues présentent un grand intérêt pour la cartographie de l’érosion côtière. The cliffs of Upper Normandy have high erosion dynamics which have mostly been studied by monitoring the cliff top. By integrating the orthophotograph of 2008 (BD Ortho®, IGN) to the initial photogrammetric analysis (1966-1995, Costa, 2000), this work provides results over a longer time period (1966-2008, so 42 years). To determine the rate of cliff top retreat, two calculation methods on fine scale (every 50 m) were tested: the “baseline approach” and the “area-based approach”. These retreat rates were calculated every 50 m in order to highlight the fine scale phenomena (especially the influence of shingle transit obstacles as major falls or port piers leading to multiplication by 3 to 4 of ablation rates). Beyond the traditional approach of estimating retreat rates, a work about retreat modalities was conducted between Etretat and Le Treport thanks to the inventory of falls on vertical aerial photographs of National Geographic Institute (IGN) (1939-1995), previously georeferenced, and orthophotographs (2000 and 2008). An inventory of falls from 2002 to 2009 (ESTRAN non-profit organization) along 37.5 km of coast provides additional information. Associated to rates, these values bring information about evolution modalities, two areas with different dynamics are observed. The former area (Antifer cape/Paluel and Dieppe/Le Treport) has low to moderate retreat rates (from 0.09 to 0.18 m/y) and is characterized by rare but large falls (median of departure area of falls of 700 m² for Fecamp/Eletot). The latter sector (Saint-Valery-en-Caux/Dieppe) has greater retreat rate (0.23 m/y) and falls are more frequent but not massive (median of departure area of falls of 200 m² for Quiberville/Dieppe). However, thanks to the Service Littoral (ESTRAN non-profit organization), weekly fall inventory from Veules-les-Roses to Le Treport (2002-2009) statistically provides some answers about the occurrence frequency of falls and their return periods (fitting to the power law). This information corroborates retreat results obtained by photo-interpretation because, again, the area of Ailly cape appears as the most dynamic with a fall over 1,000 m3 per linear km expected every 6 month whereas it is from 2 to 3 years for other sectors (Veules-les-Roses/Le Treport, Hautot/Bracquemont), with a retreat rhythm (elapsed time between two successive falls at the same place) estimated at 7 years. The obtained information is of great interest for mapping coastal erosion.

International audience; Integrated in a wide research assessing destabilizing and triggering factors to model cliff dynamic along the Dieppe’s shoreline in High Normandy, this study aims at testing boat-based mobile LiDAR capabilities by scanning 3D point clouds of the unstable coastal cliffs. Two acquisition campaigns were performed in September 2012 and September 2013, scanning (1) a 30-km-long shoreline and (2) the same test cliffs in different environmental conditions and device settings. The potentials of collected data for 3D modelling, change detection and landslide monitoring were afterward assessed. By scanning during favourable meteorological and marine conditions and close to the coast, mobile LiDAR devices are able to quickly scan a long shoreline with median point spacing up to 10cm. The acquired data are then sufficiently detailed to map geomorphological features smaller than 0.5m 2. Furthermore, our capability to detect rockfalls and erosion deposits (>m 3) is confirmed, since using the classical approach of computing differences between sequential acquisitions reveals many cliff collapses between Pourville and Quiberville and only sparse changes between Dieppe and Belleville-sur-Mer. These different change rates result from different rockfall susceptibilities. Finally, we also confirmed the capability of the boat-based mobile LiDAR technique to monitor single large changes, characterizing the Dieppe landslide geometry with two main active scarps, retrogression up to 40m and about 100,000m 3 of eroded materials.

The aim of this paper is to examine the sediment budget along the coast of Upper-Normandy, including the gravel beach (stock, input and output). The cliff coast and pebble/gravel beach, such along Upper-Normandy coast, is also considered as a system because cliff retreat induces widening of shore platforms and supplies the pebble/gravel beach with flint. In return, shore platform and pebble/gravel beach contribute to reduce intensity and action time of the waves at the foot of chalk cliff, and therefore cliff instability. However, for this type of coast, sedimentary exchanges with shoreface and offshore are rarely taken into account because of the low crosshore mobility of pebbles/gravels. However, the internal structure of pebble beaches is very often composed of a fraction proportion which can exceed 50 %. The sand fraction, presents also on rock shore platform and creating a sandy low tide terrace, seems to participate in the adjustment of these gravel beaches with wave. Theses interactions between two domains (coarse grained beach/sandy foreshore and shoreface) require to extend seaward the limits of the hydro-sedimentary system coast of Upper-Normandy.

International audience; Xynthia storm, in February 2010, has left its mark on the French coast of Atlantic Ocean and the English Channel because of its severity (47 people died and about €2 billion in damages). Consequently, the current context gives strong interest in studies about these meteorological and marine phenomena and their weather forecasting/prevention. In this paper, we studied French low coastal areas of the eastern English Channel (Normandy and Picardy coasts) because of their sensitivity to coastal flooding risk. Because of cumulative factors (exposure to strong West wind, beach sedimentary crisis, low altimetry of valleys), the French coast of the eastern English Channel is particularly interesting for analysis of this natural risk. This work has a double objective. The first objective is the presentation of our coastal flooding inventory (from 1949 to 2010) and the detection of a possible evolution about frequency and intensity of coastal flooding events perhaps linked to current climate change. The second objective is the characterization of the different meteorological, marine, and morphological parameters that end in these destructive phenomena. In matching these meteorological and marine factors with the past events inventory, we have identified the dominance of the concomitancy of well-known conditions (strong wind perpendicular to coast and high tide level), reinforced here by the passage of an atmospheric cold front during high tide. But some cases do not follow this logic. To understand them, we have followed step by step the explanation process in adopting inductive and naturalistic approaches. The aim is to highlight other factors that are involved in the process of coastal flooding. This new point of view is based on field observation and the use of different spatial and temporal scales and enables to emphasize "the preparation time" of the beach (lowering of its profile) by meteorological and marine conditions (not necessarily extreme) before the coastal flooding event.; Après le passage de la tempête Xynthia fin février 2010, dont la gravité des inondations par la mer sur le littoral atlantique et en Manche a été exceptionnelle (en France, 47 morts et près de 2 milliards d'euros de dommages), l'étude des phénomènes météo-marins à l'origine de ces franchissements et leur prévision/prévention a connu un net regain d'intérêt. Cet article porte sur les côtes basses urbanisées du littoral normano-picard qui cumule de nombreux facteurs de prédisposition (ouverture aux flux d'ouest perturbés, crise sédimentaire des plages et faible altimétrie des vallées). Pour ces trois principales raisons, ces secteurs revêtent un intérêt tout particulier pour l'analyse du risque d'inondation par la mer. L'objectif du travail est double : d'une part, présenter l'inventaire des inondations (sur la période 1949-2010) et détecter une éventuelle évolution de la fréquence et de l'intensité de ces événements, en relation possible avec le changement climatique actuel ; d'autre part, caractériser les divers paramètres météo-marins et morphologiques des plages aboutissant aux inondations par la mer. Au total, le croisement de la base de données " submersion " et des facteurs météo-marins confirme le rôle essentiel de la concomitance de conditions bien connues (fort vent d'afflux et haut niveau marin), ici renforcées par le passage d'un front froid au moment de la pleine mer. Mais quelques cas ne répondant pas à cette logique nous ont conduits à reprendre pas à pas la chaîne des facteurs explicatifs en adoptant une démarche inductive et naturaliste. Il s'agit de mettre en évidence l'action d'autres facteurs responsables des franchissements jusqu'à présent mal appréhendés. Ce nouveau regard, fondé sur l'observation de terrain et l'emboîtement des échelles spatiales et temporelles, a notamment permis de mettre en évidence le " temps de préparation " de la plage (abaissement de son profil) par des conditions météo-marines non nécessairement extrêmes, antérieures à l'inondation.

The study is carried out in collaboration with the non-profit corporation, ESTRAN. Thanks to this original partnership, we have a remarkable and unique database describing 331 coastal chalk cliff falls from 2002 to 2009 along the chalk cliff line from Sainte-Marguerite-sur-Mer to Bracquemont (Upper Normandy, France). This database has numerous advantages: a long period of observation (8 years) and a high-periodicity of survey (every week). Spatial and temporal distributions of cliff falls highlight sensitive areas as Ailly cape and that cliff falls occur all year long. However, winter is the main season of cliff falls occurrence (in fallen volume and number of falls), then autumn and spring. Summer is not a season prone to cliff falls but some massive falls might occur. The statistical approach used in this study (P.C.A.) emphasizes that cliff fall is the result of a combination of processes. Marine processes seem mainly responsible for falls in winter, then in autumn and sub-aerial weathering processes seem predominant in summer, then in spring, but falls are less numerous and less massive. Statistical approach (fitting to a power law) provides answers about return periods of falls according to fallen volumes and spatial distributions. This statistical approach used in this study is only relevant if it is linked with geomorphologic approach in order to distinguish factors that trigger the fall., L'originalité de ce travail repose sur la collaboration avec l'association ESTRAN qui, depuis 2002, recense les effondrements des falaises crayeuses du linéaire côtier entre Sainte-Marguerite-sur-Mer et Bracquemont (Haute-Normandie). Cette base de données " effondrements " (331 effondrements entre 2002 et 2009) est unique en raison de la longueur de la période d'observation (8 ans) et de l'importante périodicité des levés (hebdomadaire). La répartition spatiale et temporelle des effondrements permet de déterminer les espaces sensibles comme le secteur du Cap d'Ailly et met en évidence leur survenue toute l'année. Cependant, l'hiver est la saison la plus propice à l'érosion (aussi bien en volume effondré qu'en nombre d'effondrements), suivi de l'automne et du printemps. L'été est la période la moins propice sans être exempte d'effondrements massifs. L'analyse statistique utilisée (A.C.P.) souligne que ces mouvements sont le produit d'une combinaison de facteurs. Les facteurs marins semblent prépondérants en hiver puis à l'automne tandis qu'en été puis au printemps, les facteurs continentaux prédominent, générant des effondrements souvent moins nombreux et moins volumineux. L'analyse statistique nous permet également de proposer des éléments de réponses quant aux périodes de retour de ces mouvements selon leurs volumes et leurs répartitions spatiales (ajustement à la loi de puissance). Cette approche quantitative utilisée n'est pertinente que si elle est couplée à une approche géomorphologique afin de distinguer les facteurs responsables du déclenchement des effondrements.