Navigation – Plan du site

AccueilNumérosvol. 34/2Morphometry, distribution and Hol...

Morphometry, distribution and Holocene dating of closed depressions, called mardelles, in northeastern France

Morphométrie, distribution et datation de l’Holocène des dépressions fermées (mardelles) dans le nord-est de la France
David Etienne, Murielle Georges-Leroy, Clément Laplaige, Anne-Véronique Walter-Simonnet, Pascale Ruffaldi et Etienne Dambrine
p. 123-134

Résumés

Les dépressions fermées (CD) sont de petites formes de relief très fréquentes dans les régions de dépôts de lœss en Europe, où leur origine - qu'elle soit géologique ou humaine - est encore débattue. Dans le nord-est de la France, ces dépressions fermées sont appelées « mardelles » et sont répandues sur différents substrats géologiques et contextes paysagers actuels. Pour étudier leurs caractéristiques morphométriques et leur distribution spatiale, nous avons utilisé les résultats de deux prospections LiDAR haute résolution qui nous ont permis d’effectuer un inventaire de 1300 mardelles. Ainsi, la distribution de ces petites dépressions n’est pas homogène puisqu’elles sont beaucoup plus fréquentes dans les forêts actuelles (70 %) que dans les prairies ou les zones cultivées. En moyenne, ces mardelles sont des structures de petite taille, d’une surface moyenne de 347 m² (médiane : 449 m²), et plus de 80 % ont un diamètre de 10 à 30 m. Leur présence a été identifié sur tous substrats géologiques régionaux, et pas seulement sur les marnes du Keuper. De plus, les datations radiocarbone des sédiments les plus profonds disponibles pour 23 d’entre-elles à l'échelle locale et régionale suggèrent qu’il coexiste dans cette région des dépressions fermées d’origines différentes.

Haut de page

Texte intégral

We are grateful to all institutions that provided us with the LiDAR data, Réseau Ferré de France (RFF) and National Geographic Institute (IGN). The authors would like express their gratitude to Jean-Noël Avrillier for GIS assistance.

1 - Introduction

1Closed depressions (CDs) have been reported in numerous countries throughout the European loess belt from eastern to western Europe. These small landforms are generally not connected to the hydrological network; therefore, they accumulate direct rainfall and litterfall inputs as well as erosion products from surrounding soils (Norton, 1986; Tiner, 2003). In a review, Kołodyńska-Gawrysiak & Poesen (2017) gathered available data concerning closed depression morphometric characteristics (diameter, area and depth), their occurrence in the present landscape (forest or cropland), lithology and dating of sediment accumulations and discussed their natural or anthropogenic origins. They showed that all these small landforms may not have been formed by the same processes and may have resulted from multiple natural processes, such as (1) dissolution of calcium carbonate within the loess cover, possibly caused by deflation and uneven loess deposition concentrations; (2) dissolution of gypsum/limestone/halite lenses or along faults or cracks of tectonic or periglacial origins in the basement below the loess cover and soil subsidence (sinkholes); (3) soil subsidence caused by melting of the permafrost (lithalse, thermokarst); (4) cryokarstic hollows of periglacial origin; and (5) increased dissolution induced by water accumulation in depressions. On the other hand, they may also have resulted from human activities, such as limestone or marlstone extraction, cattle troughs, collapse of underground quarries, and bomb craters. However, they may also have resulted from mixed origins because humans may have modified naturally occurring depressions or because deforestation may have changed the water budget and soil erosion that changed their formation and infilling. Finally, they suggested that the morphological parameters and landscape distribution of these small landforms may differ in relation to their anthropogenic or natural origins.

2Closed depressions are common in Luxemburg, Belgium and northeastern France. In northeastern France (Moselle, Lorraine), large-scale inventories were performed by the forest administration at the beginning of the 20th century (Wichmann, 1903). Over 5,000 CDs, locally referred to as “mardelles” regardless of their morphometric or sediment-infilling characteristics, were identified in all geological contexts, i.e., on marlstone and dolomitic marlstone (Lower and Upper Keuper), limestone (Lias and Muschelkalk) and sandstone (Buntsandstein and Rhaetian). Their formation processes have been debated for a century (Grenier, 1906; Welter, 1908; Deffontaines & Guyot, 1922; Linckenheld, 1927; Delafosse, 1948; Slotboom, 1963; Coûteaux, 1969; Dufrêne & Legendre, 1991; Gamez et al., 2000; Barth et al., 2001; Gillijns et al., 2005; Vanwalleghem et al., 2007; Etienne et al., 2010, 2011; Ollive et al., 2021).

3Digital elevation models derived from LiDAR (light detection and ranging) surveys are increasingly used for archaeological (e.g. Georges-Leroy et al., 2012; Opitz & Cowley, 2013; Georges-Leroy & Viller, 2016; Hadjimitsis et al., 2020), ecological (e.g.,Lefsky et al., 2002; Reutebuch et al., 2005; Kissling et al., 2017) or geological (e.g.,Van Den Eeckhaut et al., 2007; Dong, 2012) purposes, but the LiDAR-derived digital elevation models available are still underexploited for inventories of geomorphological features (e.g.,Wall et al., 2016; Matos-Machado, 2018; Bensaadi & Hecker, 2019). In Lorraine (NE France), two LiDAR surveys were performed in landscapes where CDs were formerly identified. Following a suggestion by Kołodyńska-Gawrysiak & Poesen (2017), we hypothesized that their morphometric characteristics (diameter and area) and their distribution in the landscape, added to dating of sediment infilling, may help us to understand their potential formation processes. To test this hypothesis, we used both LiDAR surveys to (i) perform an exhaustive inventory of CDs in the prospected areas covering different geological basements; (ii) measure the area of CDs and the distance to the nearest CD and determine the slope of the terrain surrounding the CDs; and (iii) analyse these parameters in relation to the landscape and geological contexts. Finally, (iv) we compiled the available radiocarbon dates of sediment infilling at local and regional scales and discussed their relationship to environmental parameters.

2 - Materiel and methods

2.1 - Environmental setting

4The study area is located in Lorraine (eastern part of the Paris Basin; fig. 1A) and more specifically in Moselle (fig. 1B). The geological basement is composed of limestone (Lias and Muschelkalk), sandstone (Buntsandstein and Rhaetian), marlstone and dolomitic marlstone (Lower and Upper Keuper) (fig. 1B; BRGM, 2001). Silt deposits (0 to 2 m thick) cover the geological substrata. The landscape is composed of smooth hillslopes and flat and undulating valleys with a general southwest/northeast orientation and an elevation ranging from 150 to 300 m a.s.l. The climate is semicontinental, with an average annual rainfall of 760 mm and a mean temperature of 9.5 °C. The vegetation includes large broadleaf forests and patches of cropland and grassland.

Fig. 1: Study area. Fig. 1 : Zone d’étude.

Fig. 1: Study area. Fig. 1 : Zone d’étude.

(A) Geographical location of study area in Moselle (Lorraine, France) and of Verzy (VER), Altkirch (ALT21, ALT22), Tragny (TRA) and Rambervillers (RAM) CDs where sediments of closed depressions were cored and dated. (B) Present geological context (BRGM, 2001) and closed depression inventory of Wichmann (1903) in Moselle (Lorraine, France). (C) Location of both LiDAR prospections (LiDAR LGV and LiDAR IGN) studied in Lorraine (Moselle, France) and the windows (rectangle in dotted line) where sediments of closed depressions were cored and dated.
(A) Localisation géographique de la zone d’étude en Moselle (Lorraine, France) et des mardelles de Verzy (VER), Altkirch (ALT21, ALT22), Tragny (TRA) et Rambervillers (RAM) pour lesquelles les sédiments ont été prélevés et datés. (B) Contexte géologique actuel (BRGM, 2001) et inventaire des mardelles de Wichmann (1903) en Moselle (Lorraine, France). (C) Localisation des prospections LiDAR (LiDAR LGV et LiDAR IGN) réalisées en Lorraine et des fenêtres (rectangles en lignes discontinues) où des sédiments de mardelles ont été extraits et datés.

2.2 - LiDAR processing procedures

5The first LiDAR survey (LiDAR LGV) was performed along the high-speed railway line between Paris and Strasbourg. In March 2008, before bud break, a 1 km-large and 102 km-long strip was surveyed from Baudrecourt (Moselle) to Vendenheim (Strasbourg suburbs, Bas-Rhin) with an average emitted pulse density of 8 m-² (fig. 1C). The second survey (LiDAR IGN) was conducted between the cities of Battaville and Sarre Union by the French National Geographic Institute (IGN) in May 2009 (fig. 1C). A 35 km-long and 3.5 km-wide strip (121 km²), oriented SW-NE, was surveyed with a lower average emitted pulse density (4 m-²) and during the tree leaf period, which induced a variable soil response that was lower in mature forest stands than in young stands (fig. 2). The technical parameters (scanner type, flying height, scan angle, and average emitted pulse density: 8 m-²) for both missions are available in Georges-Leroy & Viller (2016).

Fig. 2: Satellite ortho-photography (A) and digital terrain model of (B) LiDAR LGV and (C) LiDAR IGN from the overlap forested area, illustrating the difference in resolution of the two prospections, based on shady LiDAR pictures (azimut 315° and elevation 45°).
Fig. 2 : Orthophotographie (A) et modèle numérique de terrain sur la base d'images LiDAR ombragées (azimut 315° et élévation 45°) des prospections LiDAR LGV (B) LiDAR IGN (C) de la zone de chevauchement, illustrant la différence de résolution des deux prospections LiDAR.

Fig. 2: Satellite ortho-photography (A) and digital terrain model of (B) LiDAR LGV and (C) LiDAR IGN from the overlap forested area, illustrating the difference in resolution of the two prospections, based on shady LiDAR pictures (azimut 315° and elevation 45°).Fig. 2 : Orthophotographie (A) et modèle numérique de terrain sur la base d'images LiDAR ombragées (azimut 315° et élévation 45°) des prospections LiDAR LGV (B) LiDAR IGN (C) de la zone de chevauchement, illustrant la différence de résolution des deux prospections LiDAR.

2.3 - Spatial analysis

6The identification and digitalization of closed depressions were achieved using shady LiDAR pictures with parameters that included an azimuth of 315° and elevation of 45°. The obtained inventories of CDs were compared with existing soil/land-use maps (SOeS, 2006) and geological maps (BRGM, 2001).

2.3.1 - Sizes of depressions and distance between structures

7After manually drawing the contour of each CD, areas were measured using the “calculate areas” module in ArcGis® software. The distances between CDs were based on the measurement between the two nearest central points (centroids) from which the radii of the two structures of interest were subtracted.

2.3.2 - Topographic position of CDs

8Ridge edges were drawn manually using contour lines (Contour function, step 50 cm), and valley bottoms were delineated using contour lines and the ArcHydro_9 module in ArcGis® software to trace the river system. The distance from each landform to the ridge and valley bottom was then measured automatically (Near function) for each centroid and transformed into a ratio between the distance to the ridge (m) divided by the distance from the ridge to the valley bottom (m).

2.4 - Radiocarbon dating

9Isolated plant remains (leaf, seed or charcoal) from sediment accumulations were manually collected during archaeological excavations (Lucy11, Lucy18, LAN, BIS, MB41 and MH2 CDs) or from sediment cores extracted from the central parts of CDs using a Russian peat sampler (GYK type, 50 cm long and 8 cm diameter). The dataset included 23 CDs located on different geological substrata in our study area and at a regional scale (fig. 1C). AMS radiocarbon dating (acid-alkali-acid pretreatment) was performed by Poznań Radiocarbon, and radiocarbon ages were calibrated with OXCAL 3.10 using the Intcal20 calibration curve (Reimer et al., 2020).

3 - Results

10The inventories of CDs from the LiDAR LGV and LiDAR IGN surveys resulted in the identification of 550 and 791 CDs, respectively. CDs cut by roads or located on LiDAR survey borders were discarded.

3.1 - Control of lidar survey quality

11The results of the two LiDAR surveys were compared over a 3.63 km² window where they overlapped. This area was mainly composed of mature hardwood forests (3.52 km²; fig. 2A-C). The manual inventory of CDs in this window allowed for the identification of 41 and 140 CDs from the LiDAR IGN and LiDAR LGV surveys, respectively. Hence, only 35 % of CDs were detected under mature forest using the LiDAR IGN survey and thus could not be used for their density calculation. The CD density and areas and distances between them were calculated using only the inventory based on the LiDAR LGV survey. However, we used the LiDAR IGN survey to determine the distribution of CDs in the landscape, despite its low resolution, because this survey is oriented in the same direction as the main relief, which is the opposite of the LiDAR LGV survey.

3.2 - Closed depression density

12The proportion and density of the CDs in the landscape computed from the LiDAR LGV survey was much higher in forests (69 %; 17.7 CDs km-²) than in croplands (17 %; 2.3 CDs km-²) or pastures (14 %; 2.3 CDs km-²) (fig. 3A-C). This difference was not detected from the LiDAR IGN survey in forests (71 %; 7.4 CDs km-²) compared with croplands (15 %; 6.4 CDs km-²) or pastures (14 %; 6.6 CDs km-²) because detection of CDs below forest was poor. On the other hand, the density of CDs was higher in cropland and grassland (fig. 3B-C). Considering the LiDAR LGV survey, the mean area of CDs was 447 m² (median: 357 m²) (fig. 4). The area of CDs in forests (mean: 373 m²; median: 305 m²) was close to that in grasslands (mean: 428 m²; median: 354 m²; difference not significant) (fig. 4A). The CDs in croplands (mean: 732 m²; median: 609 m²) were statistically larger than those in forests (p =< 10-2) or grasslands (p =< 10-2). The levels of significance were tested using a Wilcoxon test (Wilcoxon, 1945).

Fig. 3: Inventories of closed depressions all along the (A) LiDAR LGV and (B) LiDAR IGN strips representing the CDs density (per km²) based on a square of 500 m on a side. The table (C) summarized the proportion and density of CDs in each landscape context (forest, grassland and cropland) in each LiDAR strip.
Fig. 3 : Inventaires des mardelles le long des bandes de prospection LiDAR LGV (A) et IGN (B) représentant leur densité (par km²) sur la base d'un carré de 500 m de côté. Le tableau (C) résume la proportion et la densité des mardelles dans chaque contexte paysager (forêt, prairie et terres cultivées) pour chaque bande de prospection LiDAR.

Fig. 3: Inventories of closed depressions all along the (A) LiDAR LGV and (B) LiDAR IGN strips representing the CDs density (per km²) based on a square of 500 m on a side. The table (C) summarized the proportion and density of CDs in each landscape context (forest, grassland and cropland) in each LiDAR strip.Fig. 3 : Inventaires des mardelles le long des bandes de prospection LiDAR LGV (A) et IGN (B) représentant leur densité (par km²) sur la base d'un carré de 500 m de côté. Le tableau (C) résume la proportion et la densité des mardelles dans chaque contexte paysager (forêt, prairie et terres cultivées) pour chaque bande de prospection LiDAR.

Fig. 4: (A) Distribution of CDs area (m²) along the LiDAR LGV strip in relation with their location in present forest (2), grassland (3) and cropland (4) contexts compared to the complete inventory of CDs (1). (B) Distribution of CDs area (m²) whatever their present environmental and geological contexts. (C) Distributions of measure inter-distance between CDs presently located in forest context. Fig. 4 : (A) Superficie des mardelles (en m²) le long de la bande de prospection LiDAR LGV selon leur localisation actuelle en forêts (2), prairies (3) et champs cultivés (4) par rapport à l'inventaire complet. (B) Distribution des superficies des mardelles (m²) quel que soient leurs contextes environnementaux et géologiques actuels considérés et (C) des distances inter-mardelles pour les mardelles actuellement localisées en contexte forestier.

Fig. 4: (A) Distribution of CDs area (m²) along the LiDAR LGV strip in relation with their location in present forest (2), grassland (3) and cropland (4) contexts compared to the complete inventory of CDs (1). (B) Distribution of CDs area (m²) whatever their present environmental and geological contexts. (C) Distributions of measure inter-distance between CDs presently located in forest context. Fig. 4 : (A) Superficie des mardelles (en m²) le long de la bande de prospection LiDAR LGV selon leur localisation actuelle en forêts (2), prairies (3) et champs cultivés (4) par rapport à l'inventaire complet. (B) Distribution des superficies des mardelles (m²) quel que soient leurs contextes environnementaux et géologiques actuels considérés et (C) des distances inter-mardelles pour les mardelles actuellement localisées en contexte forestier.

The box in the plot represents the interquartile range (between the 25th and 75th percentiles of the data), the line represents the median while the square inside the box represents the mean of the data.
La boîte à moustache représente l’intervalle interquartile des données avec le bas et le haut de la boîte représentant les 25e et 75e quantiles. La ligne centrale indique la médiane des données, tandis que le carré indique la moyenne des données.

3.3 - Closed depression distances and topographic locations

13The distances between closed depressions were calculated for CDs located in forests, where the network seemed to be best preserved. The mean distance between CDs was 94 m (median: 80 m; fig. 4C). Ninety percent of the CDs were located on hill ridges and upper slopes (fig. 5A-B).

Fig. 5: (A) Distribution and (B) box-plot representation of CDs topographic location whatever their present environmental and geological context considered.
Fig. 5 : (A) Distribution et (B) représentation en boîtes à moustaches de la localisation topographique des mardelles quelle que soient leur contexte environnemental et géologique actuel considéré.

Fig. 5: (A) Distribution and (B) box-plot representation of CDs topographic location whatever their present environmental and geological context considered.Fig. 5 : (A) Distribution et (B) représentation en boîtes à moustaches de la localisation topographique des mardelles quelle que soient leur contexte environnemental et géologique actuel considéré.

3.4 - Inventories, sizes and distances between closed depressions in relation to their geological setting

14This comparison was only performed in forests where CDs were better preserved. CDs were more frequent on Lower Keuper marlstone (29.7 km-2 in forests) than on Muschelkalk limestone (15 km-2), Lias limestone (10.7 km-2), Upper Keuper dolomitic marlstone (7.3 km-2), Rhaetien sandstone (5 km-2) or other Triassic sandstones (Buntsandstein, 1.8 km-2) (tab. 1). We tested the influence of the geological substratum on CD size and distance. Unfortunately, this selection produced very low numbers of CDs in some geological contexts (Buntsandstein and Rhaetien sandstones and Upper Keuper dolomitic marlstone).

Tab. 1 - Closed depressions density (per km²) considering their geological implantation and their present environmental context location. Tab. 1 - Densité des mardelles (par km²) selon leur implantation géologique et leur localisation dans le contexte environnemental actuel.

Tab. 1 - Closed depressions density (per km²) considering their geological implantation and their present environmental context location. Tab. 1 - Densité des mardelles (par km²) selon leur implantation géologique et leur localisation dans le contexte environnemental actuel.

15The box-plot representations and statistical tests for CD areas (fig. 6A) and distances between CDs (fig. 6B) related to their existing geological location suggested that CDs can be split into two groups: 1) landforms located on Lower Keuper marlstone, Buntsandstein sandstone and Muschelkalk limestone; and 2) landforms located on Rhaetien sandstone, Lias limestone and Upper Keuper dolomitic marlstone.

Fig. 6: Box plots representation of area (A) and inter-distance (B) of closed depressions, in relation with their present geological implantation. Fig. 6 : Représentation à l’aide de boîtes à moustaches de la surface (A) et de l'inter-distance (B) des mardelles, selon leur implantation géologique actuelle.

Fig. 6: Box plots representation of area (A) and inter-distance (B) of closed depressions, in relation with their present geological implantation. Fig. 6 : Représentation à l’aide de boîtes à moustaches de la surface (A) et de l'inter-distance (B) des mardelles, selon leur implantation géologique actuelle.

The levels of significance (***: p ≤ 0.001; **: p ≤ 0.01; *: p ≤ 0.05) was tested using a Wilcoxon test. Black squares correspond to cross-correlations or already presented, and white squares to non-significant statistical correlations. The box in the plot represents the interquartile range (between the 25th and 75th percentiles of the data), the line represents the median while the square inside the box represents the mean of the data.
Les seuils de significativité (*** : p ≤ 0,001 ; ** : p ≤ 0,01 ; * : p ≤ 0,05) ont été testés à l'aide d'un test de Wilcoxon. Les carrés noirs correspondent à des corrélations croisées ou déjà présentées, et les carrés blancs à des corrélations statistiquement non significatives. La boîte à moustache représente l’intervalle interquartile des données avec le bas et le haut de la boîte représentant les 25e et 75e quantiles. La ligne centrale indique la médiane des données, tandis que le carré indique la moyenne des données.

3.5 - Radiocarbon dates of closed depressions

16Radiocarbon dates obtained from the deepest available sediment within CDs located on different geological substrata (fig. 7) suggested that infilling began during two main periods: during the beginning of the Subboreal period at approximately 5000 BP and during the Subatlantic period at approximately 2000 BP. Two CDs with sediments dated 5000 BP (Lucy 18 and Lucy 11) were located on Upper Keuper dolomitic marlstone (Goepp, 2010; Ollive et al., 2016). All other CDs located on Lias limestone (BSO and TRA), Lower Keuper marlstone (ALT22, ALT31, RAM, SAR, BIS, MH2, LAN, ROM, STJ, ALB, ASS, BAS and HON), Muschelkalk limestone (DOL), Buntsandstein sandstone (FAL1 and FAL2) and limestone and marl alternations from the Upper Eocene (VER) were dated at approximately 2000 BP.

Fig. 7: Radiocarbon dates and comparison of calibration distribution obtained on sediment from closed depressions located in the LGV survey and at a local and regional scale. Fig. 7 : Liste des datations radiocarbone et comparaison des datations calibrées réalisées sur les sédiments des mardelles dans la zone de la prospection LiDAR LGV, ainsi qu’à l'échelle locale et régionale.

Fig. 7: Radiocarbon dates and comparison of calibration distribution obtained on sediment from closed depressions located in the LGV survey and at a local and regional scale. Fig. 7 : Liste des datations radiocarbone et comparaison des datations calibrées réalisées sur les sédiments des mardelles dans la zone de la prospection LiDAR LGV, ainsi qu’à l'échelle locale et régionale.

The asterisk identifies closed depressions for which dated sediment has been picked on the field after complete excavations, while sediment for other CDs has been cored using a Russian corer. Red sites correspond to those whose base of filling is dated, whereas the orange sites do not.
Les astérisques identifient les dépressions fermées pour lesquelles les sédiments datés ont été prélevés sur le terrain après des excavations complètes, tandis que les sédiments des autres mardelles ont été carottés à l'aide d'un carottier russe. Les sites rouges correspondent à ceux dont la base de remplissage est datée, contrairement aux sites en orange.

4 - Discussion

17LiDAR surveys provide the opportunity to establish precise morphometric “identity cards” for closed depressions in Lorraine with a mean area of 347 m² (median: 449 m²). Considering that they generally have circular or ovoid shapes (Barth et al., 2001), their sizes are mainly distributed (82 %) in a range of 10 to 30 m in diameter (78 to 700 m²). Twelve percents are between 30 and 40 m in diameter (700 to 1256 m²), while only three percent are larger than 40 m or smaller than 10 m in diameter (fig. 4B). In forests, their areas are similar in Lorraine and in central Belgium; in cropland and grassland, CDs in Lorraine are smaller than those in central Belgium potentially due to a lower levelling impact on our CDs by ancient and modern agricultural practices (fig. 4); in comparison, whatever their present landscape location, CDs in eastern Europe are five times larger (Kołodyńska-Gawrysiak & Poesen, 2017) potentially due to differences in their formation processes and local geological context.

18Closed depressions in our study area are mainly located in forests (70 %) rather than in grasslands (16 %) or croplands (14 %). We do not know any process that would favour the formation of CDs in forests rather than in grassland or cropland. One hypothesis to explain this preferential location is that it is easier to identify landforms by LiDAR survey in forests and, therefore, there is an underestimation in grasslands and croplands due to sediment infilling and levelling by agriculture. However, archaeological prospecting conducted along the LiDAR LGV strip did not result in the discovery of numerous filled depressions not previously identified using the LiDAR LGV survey (Goepp, 2010). Another hypothesis can be the formation of these landforms in a nonspecific context and then a shift of the surrounding landscape to a forested context during historical times. At a short time scale, this hypothesis is not confirmed by a comparison between ancient maps (Carte d’État major, 1826-1831) and current soil occupation maps (SOeS, 2006) on either LiDAR strip. This reflects a high stability in forested areas, with a slight increase in the LiDAR IGN strip (4.6 %) and a slight decrease in the LiDAR LGV strip (5.9 %) during the last two centuries. At a larger time scale, paleoecological studies suggested the formation of the studied CDs in an open landscape mainly occupied by pastures (Etienne et al., 2011, 2015) prior to reforestation at the beginning of the Medieval period and the preservation of a forested context surrounding the studied CDs today (Etienne et al., 2013). These results confirm that these landforms were not formed in a forested context but do not clarify the reason for their main present location in forests. However, it can be noted that in the present landscape on the Lower Keuper basement, forests are preferentially located upstream and on thicker silt deposits, while cropland and grassland are located downstream, where the clay layer is close to the soil surface (Etienne, 2011; Burst et al., 2020).

19These inventories also indicated that CDs are preferentially located on topographic summits or upper slopes in accordance with observations conducted by ancient erudite locals at a regional scale and on closed depressions in the Merdaal forest (Gillijns et al., 2005; Vanwalleghem et al., 2007, 2008). Their preferential locations on the upper topographic summits seem to be inconsistent with the formation hypothesis that dissolution processes caused low water residence in this topographic context, in contrast to the lower topographic point, which is consistent with the natural area of water accumulation. In the western part of the Lorraine Plateau (Meuse, Lorraine), Jaillet (2005) differentiated two types of landforms depending on their landscape locations and geological settings. On the one hand, small landforms with undetermined origins are located on topographic summits and on compact and pliable grey clays covered by loess. On the other hand, sinkholes identified as deeper structures were observed in higher densities and arranged in lithostratigraphic contact alignments. Some alignment of CDs on the upper summit of topographic slopes were identified in our study area, with a NE-SW orientation consistent with the local topographic orientation but without contact alignment.

20In numerous studies from the beginning of the XXth century (Wichmann, 1903; Colbus, 1905; Grenier, 1906), local scholars identified CDs on each geological substratum at a regional scale. In our study area, we confirm that closed depressions locally called “mardelles” can be identified on each type of existing geological substratum (sandstone, limestone and marlstone). Statistical analyses suggested that closed depressions located on Keuper dolomitic marlstone, Lias limestone and on Rhaetien sandstone appear to be different from structures located on other geological substrata (fig. 6). Goepp (2010) studied two CDs on Keuper dolomitic marlstone (Lucy 11 and Lucy 18, fig. 1) during archaeological excavations along the LGV layout. These structures were larger (8850 and 5300 m², respectively) and deeper (470 cm and 390 cm, respectively) than CDs previously identified. Their basal sediments were either organic (Lucy 18) or inorganic (Lucy 11), while the basal sediments of the studied CDs on the Buntsandstein sandstone, Muschelkalk limestone and Keuper iridescent marl in Lorraine are always inorganic (blue-grey pliable clays) (Wichman, 1903; Grenier, 1906; Etienne et al., 2011; Etienne, 2011). Pieces of wood (Lucy 18, sampled at a depth of 250-290 cm) and organic peat (Lucy 11, sampled at a depth of 340-390 cm) were dated and confirmed that their formation occurred at least at 5000 cal. BP at Lucy 11 (4500 ± 35 BP, 5305-4994 cal. BP) and at Lucy 18 (4585 ± 40 BP, 5451-5054 cal. BP) because the basal layers had not been dated (structures were almost 470 cm deep). These structures (Lucy 11 and Lucy 18) have geomorphological characteristics, with a plunge of lateral layers and a cone-shaped profile (Goepp, 2010), and they have sizes and sediment dating that are very different from the results obtained for the studied CDs in other geological contexts (Gillijns et al., 2005; Vanwalleghem et al., 2007; Etienne et al., 2011). These geomorphological features located on Keuper dolomitic marlstone were formed during more ancient periods and definitely by natural processes. This confirms the presence, in a similar biogeographic area, of geomorphological structures with different origins that can be distinguished by their dates of formation and their morphological parameters in relation to their geological setting, as proposed by Gillijns et al. (2005). The radiocarbon dates obtained for the bottom sediments of the studied CDs located in the LiDAR LGV strip corresponded to the Subatlantic period (beginning at 2700 BP) and more specifically from the end of the second Iron Age or during Antiquity (fig. 7). These radiocarbon dates for CD sedimentation beginning in this short strip are entirely consistent with other dates obtained at a regional scale (Etienne et al., 2011, 2015).

5 - Conclusion

21In Lorraine, closed depressions are much more frequent in forests (~18 per km²), on Lower Keuper geological basement (~30 per km²), in the upper topographic position, with a mean distance between these landforms of 80 m and a diameter generally from 10 to 30 m (78 to 700 m² in area). CDs are less frequent in cropland and grassland, possibly due to having a better-preserved original form in forests or a profound modification in preferential location in agrarian/forest areas between the date of their formation and today. CDs are also present on all different geological substrata, i.e., Muschelkalk limestone (~15 per km² in forests), Lias limestone (~11 per km² in forests), Upper Keuper dolomitic marlstone (~7 per km² in forests), Rhaetien sandstone (~5 per km² in forests) and Buntsandstein sandstone (~2 per km² in forests), but are generally less frequently inventoried. Apart from CDs inventoried on Upper Keupper dolomitic marlstone that were identified as larger and with a more ancient sediment infilling (approximately 5000 cal. BP), CDs inventoried and dated (from the end of the second Iron Age or during Antiquity) in different geological and landscape contexts seem to be consistent with the global “identity card” established. Finally, we also suggest that morphometric characteristics, landscape distribution and dating of sediment infilling of CDs can be used to distinguish the different types of structures hitherto grouped under the term "closed depressions" but which potentially correspond to geomorphological landforms that have completely different origins but are present in the same geographic area.

Haut de page

Bibliographie

BARTH B., SCHNEIDER C., SCHNEIDER T.-G., DORDA D., EISINGER D., DIDION A. & ROYAR H., 2001 - Les mardelles en Sarre et en Lorraine. O.N.F, Nancy, 265 p.

BENSAADI K. & HECKER A., 2019 - LiDAR: Towards a Better Understanding of Karst Landscapes. Paysage de l’Eau, 20, 1-17.

BRGM, 2001 Carte géologique de la France à 1/50 000 (n°195), Château-Salins. Bureau de Recherches Géologiques et Minières, Orléans.

BURST M., CHAUCHARD S., DAMBRINE E., DUPOUEY J.-L. & AMIAUD B., 2020 - Distribution of soil properties along forest-grassland interfaces: Influence of permanent environmental factors or land-use after-effects? Agriculture Ecosystems & Environment, 289, 106739.

COLBUS E., 1905 - Neue Untresuchungen von Maren und der danebenliegenden Tumuli. Jahrbuch des Gesellshaftfür Lothringische Geschichte und Altertumskunde, 17, 236-271.

COÛTEAUX M., 1969 - Formation et chronologie palynologique des tufs calcaires du Luxembourg Belgo-Grand-Ducal. Bulletin Association Française d’Etude du Quaternaire, 20, 179-206.

DEFFONTAINES M.M. & GUYOT H., 1922 - Les mardelles de Lorraine. Revue. Anthropologique, 11-12, 359-362.

DELAFOSSE W., 1948 - De l'origine des Mardelles de Lorraine. Extraits des Memoires de l’Académie Nationale Metz, 11, 1-23.

DONG P., 2012 - Editorial: Applications of Light Detection and Ranging (LiDAR) in Geosciences. Journal of Geology & Geosciences, 1, 1-2.

DUFRÊNE M. & LEGENDRE P., 1991 - Geographic Structure and Potential Ecological Factors in Belgium. Journal of Biogeography, 18, 257-266.

ETIENNE D., RUFFALDI P., RITZ F. & DAMBRINE E., 2010 - Etude des variations de la végétation d’un massif forestier de la Plaine Lorraine (Moselle, France) depuis le Moyen Age. Quaternaire, 21, 317-323.

ETIENNE D., RUFFALDI P., GOEPP S., GEORGES-LEROY M., POLLIER B. & DAMBRINE E., 2011 - The origin of closed depressions in Northeastern France: A new assessment. Geomorphology, 126, 121-131.

ETIENNE D., 2011 - Les mardelles intra-forestières de Lorraine - Origines, archives paléo-environnementales, évolutions dynamiques et gestion conservatoire. Thèse de doctorat, Université de Lorraine, Nancy, 264 p.

ETIENNE D., RUFFALDI P., DUPOUEY J.-L., GEORGES-LEROY M., RITZ F. & DAMBRINE E., 2013 - Searching for ancient forests: A 2000 years history of land use in northeastern French forests deduced from the pollen compositions of closed depressions. The Holocene, 23, 678-691.

ETIENNE D., DESTAS M., LYAUTEY E., MARTI R., RUFFALDI P., GEORGES-LEROY M., DAMBRINE E. & TOPP E., 2015 - Two thousand-year reconstruction of livestock production intensity in France using sediment-archived fecal Bacteroidales and source-specific mitochondrial markers. The Holocene, 25, 1384-1393.

GAMEZ P., HARMAND D. & CARAMELLE F., 2000 - Karst et paléo-écoulements dans le Pays-Haut (Lorraine) : l’exemple du bassin versant de la vallée d’Anderny-Malavillers. Revue Géographique de l’Est, 40, 1-19.

GEORGES-LEROY M., BOCK J., DAMBRINE E. & DUPOUEY J.-L., 2012 - Apport du lidar à la connaissance de l’histoire de l’occupation du sol en forêt de Haye. ArchéoSciences, 35, 117‑129.

GEORGES-LEROY M. & VILLER S., 2016 - Les apports du lidar à l’archéologie préventive : l’exemple de la LGV Est Paris-Strasbourg en Lorraine. Archéopages, 4, 82‑91.

GILLIJNS K., POESEN J. & DECKERS J., 2005 - On the characteristics and origin of closed depressions in loess-derived soils in Europe - a case study from central Belgium. Catena, 60, 43-58.

GOEPP S., 2010 Etude des mardelles de deux secteurs de la LGV sur les marnes du Keuper. Rapport Technique INRAP, Metz, 131 p.

GRENIER A., 1906 Habitations gauloises et villa latines dans la cité des médiomatrices, étude sur le développement de la civilisation Gallo-romaine dans une province gauloise. Librairie Honoré Champion, Paris, 212 p.

HADJIMITSIS D.G., THMISTOCLEOUS K., CUCA B., AGAPIOU A., LYSANDROU V., LASAPONARA R., MASINI N. & SCHREIER G., 2020 Remote Sensing for Archaeology and Cultural Landscapes - Best Practices and Perspectives Across Europe and the Middle East. Springer Nature, Switzerland, 283 p.

JAILLET S., 2005 Le Barrois et son karst couvert. Editions Karstologia Mémoires 12, 336 p.

KISSLING W.D., SEIJMONSBERGEN A.C., FOPPEN R.P.B. & BOUTEN W., 2017 - eEcoLiDAR, eScience infrastructure for ecological applications of LiDAR point clouds: reconstructing the 3D ecosystem structure for animals at regional to continental scales. Research Ideas and Outcomes, 3, e14939.

KOLODYNSKA-GAWRYSUAK R. & POESEN J., 2017 - Closed depressions in the European loess belt - Natural or anthropogenic origin? Geomorphology, 288, 111-128.

LEFSKY M.A., COHEN W.B., PARKER G.G. & HARDING D.J., 2002 - Lidar Remote Sensing for Ecosystem Studies: Lidar, an emerging remote sensing technology that directly measures the three-dimensional distribution of plant canopies, can accurately estimate vegetation structural attributes and should be of particular interest to forest, landscape, and global ecologists. BioScience, 52, 19-30.

LINCKENHELD E., 1927 - Les mardelles de Lorraine. Bulletin de l’Assocation Philomatique Alsace-Lorraine, 7, 169-177.

MATOS-MACHADO R., de 2018 Paysages de guerre et LiDAR : de la caractérisation des polémoformes à la conservation des patrimoines naturel et culturel de la forêt domaniale de Verdun (Meuse, France). Thèse de Doctorat, Université Paris 7, 589 p.

NORTON L.D., 1986 - Erosion-sedimentation in a closed drainage basin in Northwest Indiana. Soil Science Society of America Journal, 50, 209-213.

OLLIVE V., ERNST T., FRANCK J., MEYER N., THOMASHAUSEN L. & VILLER S., 2016 - L’Érosion agricole d’un petit bassin versant lorrain depuis 2000 ans: Géoarchéologie de la LGV-Est européenne. Quaternaire, 27, 27-37.

OLLIVE V., WIETHOLD J., KLAG T. & KLAG P., 2021 - Origine, processus de formation et évolution des mardelles du Nord-Est du Bassin Parisien (France). Bulletin de la Société Géologique de France, 192, 28-46.

OPITZ R.S. & COWLEY D.C., 2013 Interpreting Archaeological Topography: Lasers, 3D Data, Observation, Visualisation, and Applications. Oxbow Books, Oxford, England, 266 p.

REIMER P., AUSTIN W., BARD E., BAYLISS A., BLACKWELL P., BRONK RAMSEY C., BUTZIN M., CHENG H., EDWARDS R., FRIEDRICH M., GROOTES P., GUILDERSON T., HAIDAS I., HEATON T., HOGG A., HUGHEN K., KROMER B., MANNING S., MUSCHELER R., PALMER J., PEARSON C., VAN DER PLICHT J., REIMER R., RICHARDS D., SCOTT E., SOUTHON J., TURNEY C., WACKER L., ADOLPHI F., BÜNTGEN U., CAPANO M., FAHRNI S., FOGTMANN-SCHULZ A., FRIEDRICH R., KÖHLER P., KUDSK S., MIYAKE F., OLSEN J., REINIG F., SAKAMOTO M., SOOKDEO A. & TALAMO S., 2020 - The IntCal20 Northern Hemisphere radiocarbon age calibration curve (0-55 cal kBP). Radiocarbon, 62, 725-757.

REUTEBUCH S.E., ANDERSEN H.-E. & McGAUGHEY R.J., 2005 - Light Detection and Ranging (LIDAR): An Emerging Tool for Multiple Resource Inventory. Journal of Forestry, 286-292.

SLOTBOOM R.T., 1963 - Comparative geomorphological and palynological investigations of the Pingos (Viviers) in The Hautes-Fagnes (Belgium) and the Mardellen in the Gutland (Luxemburg). Zeitschrift für Geomorphologie, 7, 193-231.

SOeS, 2006 Corine Land Cover France. Guide d'utilisation. Document Technique, Ministre de l'Écologie, de l'Énergie, du Développement durable et de l'Aménagement du territoire, Service de l'observation et des statistiques. 22 p.

TINER R.W., 2003 - Geographically isolated wetlands of the United States. Wetlands, 23, 494-516.

VAN DEN EECKHAUT M., POESEN J., DUSAR M., MARTENS V. & DUCHATEAU P., 2007 - Sinkhole formation above underground limestone quarries: a case study in South Limburg (Belgium). Geomorphology, 91, 19-37.

VANWALLEGHEM T., POESEN J., VITSE I., BORK H.R., DOTTERWEICH M., SCMIDTCHEN G., DECKERS J., LANG A. & MAUZ B., 2007 - Origin and evolution of closed depressions in central Belgium, European loess belt. Earth Surface Processes and Landforms, 32, 574-586.

VANWALLEGHEM T., VAN DEN EECKHAUT M., POESEN J., GOVERS G. & DECKERS J., 2008 - Spatial analysis of factors controlling the presence of closed depressions and gullies under forest: application of rare event logistic regression. Geomorphology, 95, 504-517.

WALL J., BOHNENSTIEHL D.R., WEGMANN K.W. & LEVINE N.S., 2016 - Morphometric comparisons between automated and manual karst depression inventories in Apalachicola National Forest, Florida, and Mammoth Cave National Park, Kentucky, USA. Natural Hazards, 85, 729-749.

WELTER T., 1908 - Les mares ou mardelles : habitations souterraines celtiques en Lorraine. Bulletin de la Société Préhistorique Française, 5 (1) 41-52.

WICHMANN D., 1903 - Über die Maren oder Mertel in Lothringen : mit einem Berichtüber die Ausgrabungen des Herrn Colbus in Altrip. Jahrbuch des Gesellchaft für Lothringische Geschichte une Altertumskunde, 15, 218-262.

WILCOXON F., 1945 - Individual comparisons by ranking methods. Biometrics Bulletin, 1, 80-83.

Haut de page

Table des illustrations

Titre Fig. 1: Study area. Fig. 1 : Zone d’étude.
Légende (A) Geographical location of study area in Moselle (Lorraine, France) and of Verzy (VER), Altkirch (ALT21, ALT22), Tragny (TRA) and Rambervillers (RAM) CDs where sediments of closed depressions were cored and dated. (B) Present geological context (BRGM, 2001) and closed depression inventory of Wichmann (1903) in Moselle (Lorraine, France). (C) Location of both LiDAR prospections (LiDAR LGV and LiDAR IGN) studied in Lorraine (Moselle, France) and the windows (rectangle in dotted line) where sediments of closed depressions were cored and dated.(A) Localisation géographique de la zone d’étude en Moselle (Lorraine, France) et des mardelles de Verzy (VER), Altkirch (ALT21, ALT22), Tragny (TRA) et Rambervillers (RAM) pour lesquelles les sédiments ont été prélevés et datés. (B) Contexte géologique actuel (BRGM, 2001) et inventaire des mardelles de Wichmann (1903) en Moselle (Lorraine, France). (C) Localisation des prospections LiDAR (LiDAR LGV et LiDAR IGN) réalisées en Lorraine et des fenêtres (rectangles en lignes discontinues) où des sédiments de mardelles ont été extraits et datés.
URL http://journals.openedition.org/quaternaire/docannexe/image/17809/img-1.jpg
Fichier image/jpeg, 618k
Titre Fig. 2: Satellite ortho-photography (A) and digital terrain model of (B) LiDAR LGV and (C) LiDAR IGN from the overlap forested area, illustrating the difference in resolution of the two prospections, based on shady LiDAR pictures (azimut 315° and elevation 45°).Fig. 2 : Orthophotographie (A) et modèle numérique de terrain sur la base d'images LiDAR ombragées (azimut 315° et élévation 45°) des prospections LiDAR LGV (B) LiDAR IGN (C) de la zone de chevauchement, illustrant la différence de résolution des deux prospections LiDAR.
URL http://journals.openedition.org/quaternaire/docannexe/image/17809/img-2.jpg
Fichier image/jpeg, 432k
Titre Fig. 3: Inventories of closed depressions all along the (A) LiDAR LGV and (B) LiDAR IGN strips representing the CDs density (per km²) based on a square of 500 m on a side. The table (C) summarized the proportion and density of CDs in each landscape context (forest, grassland and cropland) in each LiDAR strip.Fig. 3 : Inventaires des mardelles le long des bandes de prospection LiDAR LGV (A) et IGN (B) représentant leur densité (par km²) sur la base d'un carré de 500 m de côté. Le tableau (C) résume la proportion et la densité des mardelles dans chaque contexte paysager (forêt, prairie et terres cultivées) pour chaque bande de prospection LiDAR.
URL http://journals.openedition.org/quaternaire/docannexe/image/17809/img-3.jpg
Fichier image/jpeg, 366k
Titre Fig. 4: (A) Distribution of CDs area (m²) along the LiDAR LGV strip in relation with their location in present forest (2), grassland (3) and cropland (4) contexts compared to the complete inventory of CDs (1). (B) Distribution of CDs area (m²) whatever their present environmental and geological contexts. (C) Distributions of measure inter-distance between CDs presently located in forest context. Fig. 4 : (A) Superficie des mardelles (en m²) le long de la bande de prospection LiDAR LGV selon leur localisation actuelle en forêts (2), prairies (3) et champs cultivés (4) par rapport à l'inventaire complet. (B) Distribution des superficies des mardelles (m²) quel que soient leurs contextes environnementaux et géologiques actuels considérés et (C) des distances inter-mardelles pour les mardelles actuellement localisées en contexte forestier.
Légende The box in the plot represents the interquartile range (between the 25th and 75th percentiles of the data), the line represents the median while the square inside the box represents the mean of the data.La boîte à moustache représente l’intervalle interquartile des données avec le bas et le haut de la boîte représentant les 25e et 75e quantiles. La ligne centrale indique la médiane des données, tandis que le carré indique la moyenne des données.
URL http://journals.openedition.org/quaternaire/docannexe/image/17809/img-4.jpg
Fichier image/jpeg, 206k
Titre Fig. 5: (A) Distribution and (B) box-plot representation of CDs topographic location whatever their present environmental and geological context considered.Fig. 5 : (A) Distribution et (B) représentation en boîtes à moustaches de la localisation topographique des mardelles quelle que soient leur contexte environnemental et géologique actuel considéré.
URL http://journals.openedition.org/quaternaire/docannexe/image/17809/img-5.jpg
Fichier image/jpeg, 219k
Titre Tab. 1 - Closed depressions density (per km²) considering their geological implantation and their present environmental context location. Tab. 1 - Densité des mardelles (par km²) selon leur implantation géologique et leur localisation dans le contexte environnemental actuel.
URL http://journals.openedition.org/quaternaire/docannexe/image/17809/img-6.jpg
Fichier image/jpeg, 102k
Titre Fig. 6: Box plots representation of area (A) and inter-distance (B) of closed depressions, in relation with their present geological implantation. Fig. 6 : Représentation à l’aide de boîtes à moustaches de la surface (A) et de l'inter-distance (B) des mardelles, selon leur implantation géologique actuelle.
Légende The levels of significance (***: p ≤ 0.001; **: p ≤ 0.01; *: p ≤ 0.05) was tested using a Wilcoxon test. Black squares correspond to cross-correlations or already presented, and white squares to non-significant statistical correlations. The box in the plot represents the interquartile range (between the 25th and 75th percentiles of the data), the line represents the median while the square inside the box represents the mean of the data.Les seuils de significativité (*** : p ≤ 0,001 ; ** : p ≤ 0,01 ; * : p ≤ 0,05) ont été testés à l'aide d'un test de Wilcoxon. Les carrés noirs correspondent à des corrélations croisées ou déjà présentées, et les carrés blancs à des corrélations statistiquement non significatives. La boîte à moustache représente l’intervalle interquartile des données avec le bas et le haut de la boîte représentant les 25e et 75e quantiles. La ligne centrale indique la médiane des données, tandis que le carré indique la moyenne des données.
URL http://journals.openedition.org/quaternaire/docannexe/image/17809/img-7.jpg
Fichier image/jpeg, 261k
Titre Fig. 7: Radiocarbon dates and comparison of calibration distribution obtained on sediment from closed depressions located in the LGV survey and at a local and regional scale. Fig. 7 : Liste des datations radiocarbone et comparaison des datations calibrées réalisées sur les sédiments des mardelles dans la zone de la prospection LiDAR LGV, ainsi qu’à l'échelle locale et régionale.
Légende The asterisk identifies closed depressions for which dated sediment has been picked on the field after complete excavations, while sediment for other CDs has been cored using a Russian corer. Red sites correspond to those whose base of filling is dated, whereas the orange sites do not.Les astérisques identifient les dépressions fermées pour lesquelles les sédiments datés ont été prélevés sur le terrain après des excavations complètes, tandis que les sédiments des autres mardelles ont été carottés à l'aide d'un carottier russe. Les sites rouges correspondent à ceux dont la base de remplissage est datée, contrairement aux sites en orange.
URL http://journals.openedition.org/quaternaire/docannexe/image/17809/img-8.jpg
Fichier image/jpeg, 449k
Haut de page

Pour citer cet article

Référence papier

David Etienne, Murielle Georges-Leroy, Clément Laplaige, Anne-Véronique Walter-Simonnet, Pascale Ruffaldi et Etienne Dambrine, « Morphometry, distribution and Holocene dating of closed depressions, called mardelles, in northeastern France  »Quaternaire, vol. 34/2 | 2023, 123-134.

Référence électronique

David Etienne, Murielle Georges-Leroy, Clément Laplaige, Anne-Véronique Walter-Simonnet, Pascale Ruffaldi et Etienne Dambrine, « Morphometry, distribution and Holocene dating of closed depressions, called mardelles, in northeastern France  »Quaternaire [En ligne], vol. 34/2 | 2023, mis en ligne le 27 juin 2023, consulté le 12 février 2025. URL : http://journals.openedition.org/quaternaire/17809 ; DOI : https://doi.org/10.4000/quaternaire.17809

Haut de page

Auteurs

David Etienne

UMR INRAE 042 - CARRTEL, Univ. Savoie Mont Blanc, FR-74200 THONON-LES-BAINS. Email: david.etienne@uni-smb.fr

Articles du même auteur

Murielle Georges-Leroy

Inspection des patrimoines - Collège archéologie, 6 rue des Pyramides, FR-75041 PARIS Cedex 01. Email: murielle.leroy@culture.gouv.fr
UMR CNRS 6249 - Laboratoire Chrono-environnement, Univ. Bourgogne Franche-Comté, FR-25000 BESANÇON Cedex.

Clément Laplaige

UMR CNRS 7324 CITERES-LAT - Univ. Tours, 33-35 allée Ferdinand de Lesseps, FR-37200 TOURS. Email: clement.laplaige@univ-tours.fr

Anne-Véronique Walter-Simonnet

UMR CNRS 6249 - Laboratoire Chrono-environnement, Univ. Bourgogne Franche-Comté, FR-25000 BESANCON Cedex. Email: Anne-Veronique.Walter@univ-fcomte.fr

Articles du même auteur

Pascale Ruffaldi

UMR CNRS 6249 - Laboratoire Chrono-environnement, Univ. Bourgogne Franche-Comté, FR-25000 BESANÇON Cedex. Emails: pascale.ruffaldi@uni-fcomte.fr

Articles du même auteur

Etienne Dambrine

UMR INRAE 042 - CARRTEL, Univ. Savoie Mont Blanc, FR-74200 THONON-LES-BAINS. Emails: etienne.dambrine@inrae.fr

Articles du même auteur

Haut de page

Droits d’auteur

CC-BY-SA-4.0

Le texte seul est utilisable sous licence CC BY-SA 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search