This article is part of the SECREI project conducted by the UR 3795 - GEGENA and co-financed by the University of Reims Champagne-Ardenne and the Grand Reims. The authors would like to thank the Association des Amis du Fort de Nogent-l’Abbesse (AFNA), in particular Jean-Michel Durlet, leader of the historical committee, and the Grand-Est region for the financing of the IMPACT 14-18 research project which allowed the acquisition of LiDAR data.

1. Introduction

1Resulting from the defeat of the Franco-Prussian war (1870-1871), France decided to develop a defensive organization in order to protect its territory. It was based on the construction of hundreds fortified structures erected along the northern and eastern borders, and more specially in the northeastern part (Ministère de la guerre, 1873) (fig. 1-2).

2This defense system, commonly referred to as “Séré de Rivières” is named after its conceptor. Until the eve of the 21^st century, as a subject of study, it was neglected by the field of scientific research. Since the 1990s, its study has nevertheless been enhanced by plural approaches emanating from architecture (Hallé, 2007; Frinjs et al., 2008; Truttmann, 2017) and military history related to the structures activity period (Neuville, 1980; Ortolan, 2004; Buffetaut, 2011, 2016; Depret, 2016; Kaluzko, 2016a, 2016b; Buffetaut 2017; Depret and Diest, 2021). Their post-conflict future was considered through the lens of the humanities and social sciences. The purpose was then heritage, landscape and urbanistic (Amat and Becker, 1994; Durup de Baleine, 1998, 2001), purely heritage (Truttmann, 1994), or historical (Diest, 2022).

3The advent of airborne LiDAR, which digitally reveals microtopography, and the development of GIS tools, has led to the emergence of geographic approaches investigating the geomorphological legacy of conflicts over the past decade, i.e., “archaeogeomorphology of conflicts” (Hesse, 2014; De Matos Machado, 2018; Gheyle et al., 2018; Taborelli, 2018).

4It took over in France from the precursory works dedicated to the consequences of WW1 (World War one) in the Paris Basin, which attest that militarization and war studies are essential to the understanding of the formation of landscapes in northeastern France (Amat, 1987, 2001; Boulanger, 2001). In this framework, the fortifications of the late 19^th century have only been approached subsidiarily in relation to WW1, by the geographical notion of polemoform (Amat, 2001, 2015; De Matos Machado, 2018; Taborelli, 2018; Devos et al., 2019) and through an archaeological prism of thought (Dardignac and David, 2015). They are thus envisaged as “ante-polemoforms” participating in the formation of WW1 landforms (Taborelli, 2018). Yet, despite the inescapable imprint of the conflict, this defensive system results in a veritable polemotopography (Amat, 2001), with its characteristic forms and structures (standardized and original on each site) and significant topographic changes.

5Until now, no approach has specifically addressed this morphological and geomorphological dimension of the “Séré de Rivières” forts, even though geomorphology, geography, and geology have strongly influenced French military thought in the 19^th century (Boulanger, 2019). The very weak chronological concordance between such social events and the science of reliefs invites to consider the fortifications as objects of geomorphological study. Although they involve extremely short morphogenetic sequences (centennial or even decadal order) their morphogenic impact remains particularly prominent in the landscape, 150 years after their construction (Taborelli et al., 2017a).

6Archaeogeomorphology, first mentioned by Wandsnider (1992) and conceptualized by Thornbush (2012, 2013) is a sub-discipline of geomorphology. This contribution, which investigates the role of fortifications and warfare in shaping the landscape of the Reims region, is regarded as archaeogeomorphological, as it studies archaeological objects by mobilizing geomorphological and geographical methods (spatial analysis, morphometry, morphological classification, GIS, LiDAR) and concepts such as polemoforms, polemolandscapes, environments (Thornbush, 2012, 2013). The geomorphological influence of Man, through his military and warlike activities, is thus considered with a cultural perspective, in parallel to anthropogeomorphology, which considers it solely through the prism of its impact on natural environments (Sherman, 1989; Goudie, 1993).

7First, this article aims to concretely introduce the influence of geomorphology and, above all, to refine our understanding of the role of geomorphological conditions in the development of this network of fortifications. Their preponderance, which is expressed at all scales, can be seen first at the regional scale, by the fortified structures that were implanted on the main reliefs (mountain massifs, cuestas) in order to protect the French capital (fig. 2A-B).

8Understanding the influence of these fortified structures on geomorphology is therefore the second objective of this contribution. More specifically, the aim is to treat these fortified structures as indicators of anthropogenic forcing on modeling in the Anthropocene, during which Man, a substantial geological force (Steffen et al., 2007), is a major morphogenic agent. In this context, the forts of the fortified belt of Reims (Marne, France) constitute an evocative case study of the topographical imprint of conflicts. The city of Reims occupies a singular geographical position, at the contact of two stepped plateaus separated by the Île-de-France cuesta: the plateau formed by the Mesozoic chalk of the Champagne region and the plateaus of the Cenozoic limestones (fig. 2B-3).

9The airborne LiDAR survey of half of this fortified belt (2015) represents a highly efficient tool for the achievement of these two objectives (Hesse, 2014), and thus to enhance the understanding of the history and environmental impact of fortifications (Devos et al., 2019).

10The Nogent-l'Abbesse fort area was selected in order to gain a better understanding of the role of these fortifications in the changing topography of these complex polemolandscapes (fig. 1E). This choice is due on the one hand on the representativeness of the site's polemotopography revealed by this very high-resolution topographic data, which evolved over 70 years from its fortification to the end of the WW2 (World War two). On the other hand, it’s also justified by the close collaboration with the site administrator enabling researchers to work on site.

11The LiDAR data was exploited through an inventory of detected topographical anomalies, based on a digitizing protocol developed beforehand (Georges-Leroy, 2011; Ancelin et al., accepted a). The crossing of this microtopographic information with archival documents, current databases and field surveys highlights the multiple morphological traces of militarization, more specifically fortification, as well as warfare. This morphological discrimination and the decomposition of the current topographical palimpsest revealed by LiDAR finally enable us to identify multiple morphogenic military and polemological phases.

12Ultimately, the key aim of this contribution is to reveal the value of “Séré de Rivières”-type military fortifications as an object of geomorphological study. Such interest is important for (i) considering their environmental impacts, (ii) improving their historical and archaeological understanding (Ancelin et al., accepted a) and (iii) contributing to the management of this little-known fortified and polemological heritage considered a dark heritage (Hesse, 2014).

Fig. 3 - Simplified geological map of Mesozoic and Cenozoic substrates of the Reims sector, including Quaternary.
Fig. 3 - Carte géologique simplifiée des substrats du Mésozoïque et du Cénozoïque, dont Quaternaire, du secteur de Reims.

2. LIDAR

2.1. Parameters

13The topographic data were acquired during an airborne LiDAR campaign conducted in December 2015 by Aerodata in the north of Reims. It covers the sectors of Brimont and Berru-Nogent-l'Abbesse inside of “Séré de Rivières” fortifications are located (fig. 4A). Favorable meteorological (low cloud cover, no snow cover) and phenological (no leaves, few coniferous stands) conditions permitted to obtain a 50-cm resolution DTM (digital terrain model) interpolated from the acquisition of an average cloud of 13 “ground” points/m² (fig. 4A-B). The density map produced at the DTM operating resolution shows a minimum density of 3 points per pixel and good isotropy, particularly for the Berru-Nogent-l'Abbesse sector) (Ancelin et al., accepted b) (fig. 4A). The resulting very high spatial resolution is consistent with geomorphological (Höfle and Rutzinger, 2011) and archaeological uses (Nuninger et al., 2010).

2.2. Cartography

14The processing of the point cloud and the DTM was carried out using Arcmap 10.3 software in order to detect and map the topographic anomalies observed (Nuninger et al., 2010).

2.2.1. DTM creation and visualization processing

15The topographic modeling was obtained by interpolating the given raw point cloud rated by the provider, in order to discriminate “ground” and “overground” points. This step was performed using the “LAS to Raster dataset” tool (parameters: inverse distance weighting; natural neighbor).

16DTM was visually processed using an algorithm (fig. 4C) to generate derived images that optimize shape detection (Kokalj and Hesse, 2017). Three different indices were used with the parameters shown in Figure 4D: multidirectional shading (Devereux et al., 2008), slope index (Bennett et al., 2012), and topographic position index (TPI) (Guisan, et al., 1999).

2.2.2. Digitization protocol

17The development of a digitization protocol standardizing the mapping work allowed the assignment of cartographic standards to each type of topographic anomaly studied (Georges-Leroy et al., 2011; Chevigny et al., 2018; De Matos Machado, 2018; Ancelin and Harmand, accepted) (fig. 4E). These are previously defined morphologically thanks to a plan view and a topographic profile. In parallel, different thematic (type, dating, etc.) and morphometric attribute fields were assigned to the created vector data table.

2.2.3. Documentary crossing and field prospection

18Finally, the protocol was verified against field observations, current documents (e.g., cartographic databases of the Institut National de l'Information Géographique et Forestière, IGN) and archival sources in manuscript, cartographic and photographic form (orthophotographs, vertical and oblique aerial photographs). The old maps and plans used date back to the 19^th century and the WW1 (tab. 1), during which millions of artillery shooting plans were developed and disseminated by the GCTA (Groupe de Canevas de Tir des Armées) of the French Army (Lévy, 1926; Combaud et al., 2016).

19The planimetric documents were georeferenced on Arcmap 10.3 with the “1^st degree polynomial” transformation method to project them simultaneously to the LiDAR data and digitized layers. This allows to constitute a dedicated, complete, and operational GIS.

3. Results

3.1. A rational and multiscalar use of the relief

3.1.1. Defensive system scale

20The influence of geomorphology on French military thought is unequivocal in view of the flagrant rationalization of the relief and its characteristics in the elaboration of the “Séré de Rivières” system (fig. 2A-B).

21First, on a national scale, the fortified structures were established on the four mountain ranges that delimited the eastern border of France (fig. 2A): on the Alpine Massif facing Italy, on the Jura Massif facing Switzerland, on the western slope of the Vosges Massif facing the new border imposed by Germany under the Treaty of Frankfurt in 1871, and on the Ardenne Massif facing Belgium.

22The northeast quarter is crucial because it is located near the German Alsace-Lorraine (Piette, 1874). It constitutes the main defense axis organized in its depth into three major defensive lines (fig. 2B). These lines are composed of a network of cities fortified by a set of detached “belt forts” of 6 km long that surround it (fig. 1A-E). Four fortified belts are in particular linked by a string of “screening” forts (Verdun-Toul and Épinal-Belfort curtains). They were supposed to generate impassable barriers, behind which the deployed troops could encircle opposing armies rushing into the military and topographical gaps located in the intervals (fig. 2B). Finally, isolated “stop” forts were placed at certain points to slow down a possible enemy invasion or to protect militarily sensitive logistic nodes.

23In addition to the mountainous massifs, the cuestas of the Parisian Basin served as support for the defensive system (fig. 2B). In the east, the 1^st line was positioned on the Meuse cuesta, covered by Oxfordian limestone (e.g., Toul-Verdun curtain) (fig. 1A, D), and rooted on the slopes of the upper Moselle (e.g., Épinal-Belfort curtain) (fig. 1B-C). Further back, the second line of defense consisted of the strongholds of Reims (fig. 3), Laon and La Fère, supported by the Île-de-France cuesta, and the strongholds of Langres and Dijon, located further to the southeast, on the eastern edge of the Langres plateau. The third line corresponds to a large francilian belt defending Paris. It was composed of fortifications scattered on the heights of the plateaus in the tertiary center of the Paris Basin.

3.1.2. Reims region (study area)

24The reliefs associated with the cuesta of Île-de-France served as the foundation for the fortified belt of Reims, which consisted of a group of 7 forts and 5 secondary fortifications (fig. 3). They are located at a good distance from the city of Reims (protection against enemy artillery) and close enough to each other to benefit from mutual lateral coverage (flanking). This combination of fortifications was placed on the highest points along the funnel of Vesle consequent valley, in order to both protect the town of Reims and to lock down this strategic route to Paris (Lacroix, 1909) (fig. 3, 5).

25Figures 3 and 5 show the geomorphological conditions which make it possible to discriminate them in 4 groups (fortifications are named from a to j). On the one hand, to the west, the Chenay redoubt (a) and the Saint-Thierry fort (b) are located on the Massif de Saint-Thierry (Saint-Thierry group). They are built to the limestone reverse side of the Île-de-France cuesta, formed by the plateaus that mark the beginning of the Cenozoic formations in the east of the Paris Basin.

26To the northwest, the fort of Brimont (d) is located on the Ypresian clays and the sand of the Thanetian. It overlooks the plain atop the eponymous mound. Its use for the defense of Reims is theoretically combined with two associated fortifications, i.e., Brimont group (Lacroix, 1909): the Loivre battery (c) located on the chalky flank (Campanian chalk) to the west; the Cran of Brimont battery (e) on the sandy flank of the mound to the east.

27To the east, most of the fortifications are built against the different geological layers of the Berru residual hill and the Nogent-l'Abbesse outlier (Berru-Nogent-l'Abbesse group). From north to south, the main line of defense is constituted by the fort of Witry-lès-Reims (g: marly sands of the Thanetian, possibly stripped during the construction of the fort), the battery of Berru (h: marls, clays and lignite sands of the lower Ypresian), and the Nogent-l'Abbesse fort and its batteries (j: continental clays and marls of the Thanetian, purplish marls and sands of the Lower Ypresian), of which the southern battery is located on the Campanian chalk. Behind this line, the vigie of Berru (i) is located on the marls (Lutetian-Bartonian) which cover the residual hill of Berru.

28The last group corresponds to the three forts positioned on the lower chalk ridges, defending the intervals located on either side of the Mont de Berru. To the north, the fort of Fresnes-lès-Reims (f) controls the interval with the Mont de Brimont. To the south, the Pompelle fort (k), dominating the Vesle valley, and the fort of Montbré (l: Mont Thibé), locked the gap with the Montagne de Reims. At the site of the latter fort, the hill is covered by sands with flint pebbles (Thanetian) dismantled during its construction (Laurain et al., 1981). Topographic (inclined plane between the plateau and the Vesle valley) and sedimentological (heterometric detrital deposits of solifuxion flows) evidence suggests that Mont Thibé is the remnant of an ancient glacis connecting the Montagne de Reims plateau to the Champagne plateau (Wiedemann, 1976) (fig. 5).

29The geomorphological description of the relief underlying the Reims belt shows that the same geomorphological situation conceals a plurality of sites. In this case, the local geological history is characterized by a shoreline situation during the upper Paleogene from Thanetian (Laurain and Meyer, 1986) to Lutetian (Gély, 1996, 2009, 2016). This situation is conducive to the existence of strong lateral variations in facies and power (fig. 5). This implies a contrasting sedimentation from one site to another, sometimes marine, tidal, fluvioestuarine, continental or lagoonal (Laurain and Meyer, 1986). This is also expressed in the variability of altitudes between the Mont de Berru, culminating at 260 m NGF (Nivellement Général de la France - General Levelling of France), and the Mont de Brimont reaching 170 m. These geological disparities induce relatively heterogeneous local availability of construction geomaterials and supply potential, despite the standardized use of certain materials (Fronteau et al., 2014; Laratte, 2022). The marine Lutetian (coarse limestone) capping the Tertiary plateaus in north of Reims, is exploited in situ for the construction of the Saint- Thierry fort and the Chenay redoubt, whereas it is absent 5 km further east on the Brimont residual hill, where the fort is based on the Thanetian gresified sands. On the other hand, in the Montagne de Reims and on the residual hill of Mont de Berru, where the eponymous vigie is located, the continental Lutetian is frankly clayey and marly (white clays and green marls on Figure 3) (Laurain et al., 1981). Similarly, water supply strategies vary with hydrogeologic conditions (Houdaille, 1889; Devos et al., 2019). Catchments from the chalk aquifer (e.g., Fort of Witry-lès-Reims or Fort of Montbré) and from the aquifer of the upper Ypresian sands - lower Ypresian clays contact (e.g., Fort of Saint-Thierry) are conducted with wells that can reach 90 m deep. In the east, the unsuitability of water consumption (decomposition of the pyrites of the Lutetian black clays) obliges the use of rainwater recovery systems (Houdaille, 1889).

3.2. The fort and battery of Nogent-l’Abbesse

30The LiDAR acquisition provides both quantitative microtopographic data and a synoptic view that are valuable in understanding the geomorphological characteristics at the scale of the Nogent-l'Abbesse fort and its detached batteries. Figure 6 shows the strong morphogenic impact of the different polemological phases that shaped this landscape. The detection of characteristic shapes (e.g., ditches, shell craters), and even certain models (WW1 landforms, or even the “Séré de Rivières” landforms), assessed by field surveys, allows us to distinguish five phases:

• During their active period, (i) their construction (fig. 6A), and (ii) their successive modernization to the torpedo shell crisis (1885-1886) (fig. 6B) before their decommissioning in 1913.

• During WW1, (iii) their integration into defense networks (fig. 6C) and (iv) their obliteration by shellfire (Devos et al., 2019) (fig. 6D).

• On the eve of WW2, (v) the construction of buried ammunition bunkers by the French army, partially destroy (distal part) by German troops in 1944 (Frinjs et al., 2008; Coutant et al., 2017) (fig. 6E).

3.2.1. Construction of the fortifications

31Whatever its name and size, a first-generation “Séré de Rivières” fortification (“1874 type”; i.e., built before 1886) corresponds to a set of structures, with or without masonry, covered by a mass of earthy materials, and dedicated directly to the artillery (firing platforms, sheltering traverses, powder, and ammunition magazines) or indirectly (barracks, soldiers living quarters, defense installations) (fig. 6A). The faces most exposed to the enemy (“head front”) are more protected than those facing the rear (“gorge front”, which have lower parapets and lesser shelters) according to the principle of irreversibility of fortification (Truttmann, 2017).

32The fort of Nogent-l'Abbesse was built between 1875 and 1879 over an area of around 16 ha. It was part of the first section of forts of the Reims belt at the end of the 1870s (Ministère de la Guerre, 1880). It has a central massif, “low battery” and is dominated by an infantry ridge (unlike the “high battery” forts known as “à cavalier”; Truttmann 2017). It is a polygonal fort whose gorge faces are pseudo-bastioned by the presence of a ravelin. Its concave pentagonal shape (the gorge faces draw a reentrant favoring flanking) corresponds to a trapezoid with a broken base. It differs from pentagonal forts whose convex head faces imply at least the presence of an additional caponier (cf. infra; e.g., Forts of Brimont and Witry-lès-Reims).

33The courtyards of the fort (“rampart street”) are punctuated by a succession of ten sheltering traverses located on the rampart, behind the scarp, designed to protect the artillery platforms from enfilade fire. Half of them are said to be “rooted” because they are extended to the parados and connected to the barracks (Frinjs et al., 2008). In addition to these lateral protections, the platforms have a frontal parapet. At its summit, the topographic ridge materializes the fire ridge behind which the cannons and even the infantry are arranged (fig. 7A).

34The fort is defended by a complete peripheral dry ditch, about 900 m long and 8 m wide at the base, and several meters deep (Houdaille, 1889) (fig. 7A-C). It is originally circumscribed laterally by two masonry retaining walls: one on the scarp side (the slope at the foot of the rampart that faces the enemy) with discharging arches, and the other on the counterscarp side (facing the fort), at the top of which is the covered way. In addition to the defensive objective, it is designed to limit the risk of clay soil landslides by optimizing drainage (longitudinal and transversal drainage profiles, presence of a cunette in the center) and creating discharge arches in the walls.

35Comparison of Figures 7A-B reveals the attenuation of its profile under the action of colluvium and ground movements (e.g., solifuxion) that are inscribed in the clayey-marl layer of the lower Ypresian (Ministère de la Guerre, 1904, 1909; Bollot, 2014), as well as by the dismantling of the walls. Indeed, after the WW1, most of this masonry was removed due to probable reuse during Reconstruction and for the construction of ammunition bunkers by the French Army in the late 1930s (Coutant et al., 2017).

36The close protection is completed by the presence of two caponiers. The first one, located in the north, is called “simple”, and the second, in the south, is called “double”. They flank respectively one and two sections of the ditch thanks to firing battlements. The diamond ditch, a small subrectangular ditch of several meters in depth, is located below the caponiers to prevent access and to limit the risk of blocking the battlements in the event of a masonry collapse (Hallé, 2007). Three annexed batteries, located in the outer perimeter, supplement the fortifications of the Nogent-l'Abbesse hillock (fig. 6A). They include a rooted sheltering traverse accompanied by one (north and east batteries) or three standard sheltering traverses (south battery). The main differences with the fort are the absence of close defense works (caponiers) and the opening of the gorge front (Frinjs et al., 2008).

37Finally, a military glacis completes the defense of the fortified complex (fig. 6A, 7A-B). It links the fort to the surrounding countryside. Its soft slope, visible from the fire ridge, should prevent the enemy from hiding from the view of the defenders. The planting of vegetation optimizes the fortification by camouflaging the massif, consolidating the slopes, and slowing down enemy attacks (tracing root system complicating the digging of offensive trenches) (Simoutre, 1889). In return, it hides the external covered way and the escarpment from the outside (Hallé, 2007).

38The absence of precise maps of the initial topographic state of the site, before and after construction (to the knowledge of the authors), complicates the quantitative estimation of the volumes disturbed, in particular for the filled central massif which necessarily involve external inputs (Simoutre, 1889). Nevertheless, Simoutre (1889) and Guillot (1949) express an initial depth of 7 m between the theoretical ground level and the bottom of the ditch. Cross-checking with LiDAR data (900 m in length, 8 m in average width) (fig. 7B-C) indicates a volume of more than 50,000 m³ initially excavated for the digging of the ditch, testifying to significant pedoturbation on the site.

3.2.2. Torpedo shell crisis and forts modification

39The discovery of smokeless powder (1884), melinite (1885) and ogival cylindrical shells with “breaking” explosives (1886) neutralized the resistance capacity of all the “Séré de Rivières” fortifications (torpedo shell crisis) (Rocolle, 1973). Their obsolescence was confirmed by the destruction tests carried out on the Malmaison fort (1886) (Ministère de la Guerre, 1886). This caused their modernization to varying degrees between the 1^st and 2^nd lines. Besides to the increasing use of narrow railroads at the end of the 19^th century, the modifications of the fort of Nogent-l'Abbesse were limited to the dispersion of the powder magazines. An extension to the latter was built in 1892 within a cavernous magazine (or magazine under rock). It was dug in the chalk underlying the glacis of the gorge front, for a total excavated volume of 989 m³ (fig. 6B, 9A). Its plan is almost identical to the cavernous magazine dug under the Vigie of Berru in the same spirit of relocation of munitions. The shelter is made up of round-vaulted rooms and is structured around two parallel access tunnels that lead to a transverse gallery. It served seven rooms, of which only one or two were devoted to housing ammunition and gunpowder (Ministère de la Guerre, 1893; Frinjs et al., 2008). Only the access platform of a little more than 1,000 m² terraced for the occasion, is visible thanks to LiDAR (fig. 6B, 9A).

3.2.3. Integration of the fortifications into the WW1 defense networks

40On the eve of the WW1, the progress of artillery led to the decommissioning of the second line forts which had become obsolete. Reims was declared an open city on September 2, 1914, causing an exodus of the population. The city was then occupied by the German army between September 4 and 12, 1914.

41After the movement war, the front settled in September 1914 between the Vesle and Aisne valleys to the north, at the base of the Île-de-France cuesta, passing through the Pompelle fort, the chalky ridges to the north-east of Reims, the Aisne-Marne canal, and the Loivre valley (fig. 3). It gradually organized itself into trench elements, then into several positions of 2 to 4 trench lines as it became denser and thicker over a depth of 20 km (Devos et al., 2015) (fig. 8A). Over the 110 km of the Champagne front, there are approximately 13,000 km of trenches organized into 4 positions (Taborelli, 2018). This network is particularly dense and thick in the Reims sector. It offers good potential for observation allowing for artillery adjustment, with numerous firing ridges below. Moreover, the Champagne chalk constitutes a substratum that can be easily dug in an unsaturated zone. The German network integrates a major part of the fortifications of the Reims belt. It contains the fortifications of Nogent-l'Abbesse (with annexed batteries), Berru (vigie, battery and under rock magazine), Witry-lès-Reims, Fresnes-lès-Reims, and Brimont (including its Cran of Brimont and Loivre batteries). The French network had only four, with the Saint-Thierry fort, the Chenay redoubt, the Montbré fort and, after a fierce battle, the Pompelle fort (the only fort retaken by French troops) (fig. 8A-B). Each network therefore benefited from exceptional observation sites such as the top of the cuesta for the French, the residual hills, and the outliers for the Germans (fig. 3). This explains the remoteness of the defense lines and the thickening of the device to protect against the more effective artillery fire in this sector (Taborelli et al., 2018) (fig. 8A).

42The forts were then integrated into the defense network, but their role depended on the position in which they were located. In the 1^st position (e.g., forts of La Pompelle and Brimont) (Devos et al., 2019; Ancelin et al., accepted a), they serve as resistance support points and land observation sites. They constitute “resistance centers” that structure the defense in cells rather than lines (Devos et al., 2019) (fig. 8B). The trench networks are organized around and below the forts following a military ridge position and no longer topographic, which allows for firing towards the front zone, but also in the intervals of the defense network. Communication trenches and tunnels allowed access to the forts. This generates on these sites typical WW1 polemoforms, but also important pedoturbations with numerous stratigraphic anomalies in the Champagne chalk (Amat, 2001; Boulanger, 2001; De Matos Machado, 2018; Taborelli, 2018; Devos et al., 2021). Moreover, on the 2^nd position, the bastion forts were too exposed to artillery shelling all over their military glacis at the top of the topographic ridge. Therefore, they were more used as observation sites and depots (cisterns, shelters) than as combat or artillery concentrations, which was the primary vocation of these works. The artillery was dispersed, generally camouflaged, behind the relief that supported the forts, to protect itself from the enemy's view. This is the case at Brimont (Devos et al., 2019) and at Nogent-l'Abbesse (fig. 8C), where the natural topography is preferentially used to advantage.

43Most of the forts, initially established to support assaults from the east, were not well exposed in relation to the front line. At Nogent-l'Abbesse, its head structures covered the east of the German position, while the front line was located 4 km to the south (fig. 8A-B). Its gorge entrance opened up to the west, facing the French positions at Reims. The access to the fort through the gorge proved to be dangerous because it became vulnerable to artillery. It required a morphological adaptation of the forts to integrate them into the defense network. Communication trenches were dug through the counterscarp of the northern battery and the head battery, as well as in the northern scarp of the fort. They then provide access to the ditch that was probably used as a traffic trench (fig. 8C). Two wooden footbridges, seen on old photographs and since disappeared, were built to cross the northern battery and the fort. The piercing of the double caponier opened the access to the interior of the fort, but it was targeted by artillery as evidenced by the shell craters revealed by LiDAR (fig. 6D). To avoid overdraft, underground “proof” galleries were dug into the chalk from the double caponier to access the outside of the fort (fig. 8C, 9B). One gallery served the 1892 powder magazine, where the Germans established a command post and an electrical plant (fig. 9A). It then continues through a tunnel to the south battery, the side of which is pierced with concrete bombproof observatories. An access tunnel crossing the rampart of this battery also links the ditch to a firing platform. The digging of these tunnels generates morphogenic rubbles that are partially transported on a narrow-gauge railroad. They were deposited in various parts of the fort, including the ditch of the south battery, in order to hide the work from the French (fig. 6C). LiDAR precisely reveals an elevation of about 3 m there corresponding to a topographic relic of the embankment, then being removed in 2015 by the site owner (Association des Amis du Fort de Nogent-l'Abbesse). The total development of the 1 km known and mapped gallery networks at the Nogent-l'Abbesse fort is accompanied by aeration/ventilation works (boreholes) and emergency exits also morphogenic. This adds up to a total excavated volume of approximately 5 000 m³.

3.2.4. Shelling and bombturbation

44The morphogenic impact of the WW1 on the forts also concerns bombturbation (Hupy and Schaetzl, 2006). The forts were targeted by enemy artillery, creating shell funnels, roof collapses, and major geotechnical disorders. The visualization processing of LiDAR DTM (slope, TPI) on the site of Nogent-l'Abbesse, counts 490 shell holes of which 330 impacts on the fort. They correspond to respective densities of 14 and 25 impacts/ha (fig. 6D). This testifies to an exacerbated bombardment on the forts, linked to destructive targeted fire (high explosive shells). The impacts (3.4 % of the spatial extent of the fortified structures) are concentrated more particularly on the barracks, the double caponier, the batteries, and in the south gallery. The distribution diagram of shell crater diameters is more or less identical to those observed elsewhere on the western front, but is more spread out, by larger diameters correlating with large-caliber shrapnel projectiles (340 to 400 mm) (fig. 6D, 9C).

3.2.5. Post-WW1 changes

45The polemological impact after the WW1 concerns the implantation of 27 underground ammunition storage cells (linked to the Air Force Base 112) (fig. 9D) on the eve of the WW2 (1935-1939) (fig. 6E). These were built on the northern flank of the mound, below the gorge front glacis. Corridors linking the bunkers three by three, and cells placed opposite the entrance serve to break the flow of explosions in case of bombardment.

46In 1944, German troops mined the 15 eastern bunkers, three of which were later rebuilt (Frinjs et al., 2008; Coutant et al., 2017) (fig. 6E). This destruction generated a 100 m diameter crater, and correlative chalky ejecta, still clearly visible in aerial photography in 1949 (fig. 9E). The LIDAR slope visualization shows that the crater was reshaped after the WW2 using leveling of the slopes or terracing operations (fig. 6E).

4. Discussion

47The geomorphological approach of the Nogent-L'Abbesse fort brings many observations such as the serial, modular and ubiquitous architecture, as well as the reasoned use of geomaterials which deserve to be relativized in space and in time.

4.1. The serial architecture of the “Séré de Rivières” system

48Before the emergence of this contrast between progressively set aside and modernized fortifications, the use of the same general principles generated a serial architecture on the defensive system scale (Diest, 2022), associated with typical morphologies. The same components (central barracks, firing platforms separated by sheltering traverses, ditch defended by caponiers, etc.) are thus reproduced in the entire Reims belt (e.g., Brimont fort) (Devos et al., 2019) or in the 1st line strongholds (e.g., Verdun and Toul belts) (Durup de Baleine, 1998, 2001). The construction phase of the Nogent-l'Abbesse fort, the most morphogenic, thus testifies to the internal organization standardization of structures prior to 1885-1886 (fig. 6A). The construction of such structures implies a topographic reworking impactful enough that geological layers are stripped (cf. supra, Fort of Montbré) (Laurain et al., 1981).

4.2. The modular and opportunistic architecture of the “Séré de Rivières” system

49A number of particularities nevertheless reflect the modular and opportunistic or ubiquitous character of these structures. To varying degrees, each fort has both a form and construction materials adapted to the topographical and geological context. The structure of the Nogent-l'Abbesse fort is remarkably adapted to the shape of the interfluve (fig. 6A, 8C). Located on the summit of the eponymous mound, it is framed by annex batteries that follow the contours of the ridge. Similarly, the annex batteries of the Brimont fort and the subtriangular shape of the Saint-Thierry fort respond to the topographic setting of their interfluve, respectively located on the flanks of the Brimont hill (chalk ridge, structural bench in the Thanetian calcarenites) (Devos et al., 2019) and at the top of the reverse side of the Île-de-France cuesta (fig. 3). However, this adaptation needs to be tempered by the ability of military engineers to shape construction sites (Ancelin et al., accepted a).

4.3. The use of geomaterials of the “Séré de Rivières” system

50Moreover, the influence of geological availability is expressed in the geomaterials used for their construction. They come from quarries approved by the Engineering Department and are generally of the same nature within the same fortified belt. The fortress of Lille was built largely of terracotta brick (Flanders clay) and the forts of the Meuse were built of Oxfordian limestone (like entroquite, Euville-Lérouville type stone) (Mégnien, 1980). The stronghold of Épinal, for its part, is partially built in sandstone of Buntsandstein (Trias). For the forts of Reims, the blue limestones of the Ardenne (Devonian), the oolithic limestones of the Meuse (Oxfordian), the Brie millstones, the Lutetian coarse limestone, the Thanetian calcarenites are used together in the form of rubble, ashlar, or paving stones. Their use is associated with the materials necessary for the manufacture of mortars and architectural terracotta (limestone, sand, and clay). This great diversity of geomaterials results from numerous architectural uses (base, elevation, framing, covering, etc.) related to the petrographic characteristics of the rocks (because rationalized). It also results from the enlargement of the supply perimeters induced by the railroad development since the middle of the 19^th century (Fronteau et al., 2014).

4.4. Polemoforms transformed by diversified trajectories

51During and after WW1, the heterogeneity of the military-civilian trajectories of the edifices implies a range of evolution of the buildings and thus, of degrees of conservation. This is appreciated both locally and in relation to distant strongholds. Indeed, the level of integration into the defense network and the shell craters density listed differ between the forts in 2^nd (e.g., the fort of Nogent-l'Abbesse) and 3^rd positions (e.g., the fort of Montbré that emerged almost intact from the conflict) (fig. 8A-C). For those located in 1^st positions (e.g., Pompelle Fort) (Neuville, 1980, 1999), Verdun (Durup de Baleine, 2001; De Matos Machado, 2018), the bombturbation is such that it generates orange peel landforms. The deluge is maximal, as under fire from trench, field, heavy high-powered and heavy railway (Perarnau et al., 2022). Firing is also diverse, ranging from barrage fire, destruction fire, counter-battery fire, or encagement fire. For those further from the front, the guns range becomes selective and less precise with targeted destruction shots of larger calibers and breaking shells. The LiDAR visualization processing in the Montagne de Reims sector (Perarnau et al., 2022), and the artillery fire plans of the GCTA effectively testifies to complex and concentrated railroads networks in firing epis for the railway artillery intended to strike the tunnels and forts occupied by the Germans. The craters density at the fort of Nogent-l'Abbesse (25 impacts/ha) (fig. 5D) is therefore logically less important than in the 1^st position such as in the Argonne (Brenot et al., 2017), or on cote 108 (Taborelli et al., 2017b), with 55 and 54 impacts/ha respectively. However, it remains high for the Reims sector, especially in 2^nd position (Devos et al., 2021).

52These facts highlight the polemological tropism of these pre-conflict fortifications, namely their propensity to concentrate shell impacts and military developments (trenches, shelters, artillery positions, observatories, etc.). These results nuance the polemological gradient usually observed at the scale of the front zone (De Matos Machado, 2018). Indeed, the latter, which is far from linear, remains high in the crossing of the defensive positions in which the “Séré de Rivières” forts are located (Taborelli, 2018). At the scale of the fortified sites, a local gradient can be observed due to the decrease in the polemoforms density as one moves away from the centers of resistance that the forts represent (fig. 8C).

4.5. Morphogenic polemoforms on a regional scale

53Furthermore, in parallel to this spatial concentration on site, the forts condition the existence of ex situ morphologies, linked or not to the WW1. This is the case of the artillery positions (mobile, fixed, railway) positioned, beyond the 4^th defense position, in the Montagne de Reims. Their presence induces the existence of blast shields, firing platforms, annexed shelters and firing epis (Perarnau et al., 2022). This spatial influence also occurs due to multiscalar geomaterial supply areas. Their perimeter implies extraction sites located near the construction site (in situ, adjacent; e.g., Thanetian calcarenite quarries of the Brimont hill), within a local radius (Lutetian limestone quarries in the Saint-Thierry Massif, millstone extraction pits in the Montagne de Reims), regional for Givetian limestone up to extra-regional with Meuse limestone.

4.6. Underground morphogenic impact

54The evolution of the war landscape of the Nogent-l'Abbesse fort implies also a morphogenic impact underground (post-1886 modifications, underground works, ammunition bunkers) (fig. 6B-C, E, 9A-B, D). Like the exo-polemoforms, hardly detected until the emergence of the new numeric tools (LiDAR, GIS), this endogenous component constitutes a type of visible landscape whose perception is made possible by field prospecting or geophysical detection techniques (Wieber, 1995). In the context of underground warfare (Triolet et Triolet, 2011), all bombproof work, done to protect from shell and camouflage, is far less morphogenic on the surface. The excavated material is typically disposed of in the forts ditches (fig. 6C). As such, these underground works preserve the topographic surface from anthropogenic erosion and induce a “polemological immunity”, comparable to the karstic immunity of the relief generated by the disappearance at depth of the hydrographic network (Salomon, 2000). This “subterranological” concept is a precursor because it will be more used in the Maginot Line in 20^th century with essentially underground structures serving “combat blocks” much less morphogenic than for the “Séré de Rivières” system (Truttmann, 2009).

5. Conclusion

55This contribution demonstrates the fundamentally geomorphological nature of the “Séré de Rivières” fortification as an object of study that deserves its place in the French geomorphology and geosciences, and also within the IAMS (International Association for Military Science). The results were obtained by an exhaustive mapping of war-related topographical anomalies appearing on the LiDAR data (fig. 4A-E). Spatial, diachronic, multiscalar and multi-criteria approaches show that this current LiDAR pattern is the result of a polyphase morphogenic process.

56On the one hand, the consideration and rationalization of the relief in the “Séré de Rivières” system is multiscalar, on the national (fig. 2), stronghold (fig. 3, 5) and fortified site scales (fig. 6). The Reims belt example is unequivocal considering the reliefs associated with the Île-de-France cuesta, which served as support for the fortified belt blocking the consequent valley funnel of the Vesle river (fig. 3, 5). At construction site scale, the forts are original polemoforms because they are paradoxically standardized in their organization but can be modulated according to varied geomorphological sites. The latter give them potential for observation and artillery fire, with diversified water and geomaterial resources.

57On the other hand, this study diachronically documents the geomorphological impacts associated with militarization and war. In the Nogent-l'Abbesse area, the various military and polemological components of the topographical palimpsest composing the landscape have been morphologically discriminated. The establishment of the fort led to a topographical disturbance associated with its construction (excavation of at least 50,000 m³ for the perimeter ditch alone) (fig. 7A-C) and modernization (1,000 m³ excavated) (fig. 6B). This underlines the impact of this pre-WWI polemoform, which tends to be overshadowed by this fundamentally morphogenic conflict, even though it partly determines its imprint on the landscape.

58One perspective of these results is anthropogeomorphological. Within the scope of the development of approaches centered on the human impact on the reliefs i.e., anthropogeomorphology (Golomb and Eder, 1964; Bhunia et al., 2022; archaeogeomorphology; Thornbush, 2012), the polemological landforms may also constitute a morphological marker of the human footprint in the terrestrial ecosystem, or even contribute to the Anthropocene definition, as proposed by Bonneuil and Fressoz (2013). While the earliest forms of pre-Neolithic warfare are only inventoried in the form of archeological material in sedimentary records (Rune, 2016) or Neolithic mass graves (Erdal and Erdal, 2012), military landforms is more fully expressed, particularly with defense morphologies (Hamilton Thompson, 1975; Keegan, 2000).

59The main perspective of this research is archaeogeomorphological. The geomorphological approach to polemolandscapes focuses on the forms, formations and processes involved in their genesis and destruction. It is tempting to apply the paradigms of geomorphological evolution of landforms to polemoforms, no longer on a geological or Pleistocene scale, but on an Anthropocene scale (protohistoric and historical epochs), with morphologies that are:

• “primitive” (remarkably well-preserved, generally simple trajectories, great potential for enhancement; e.g. Fort de Condé converted into a museum) ;

• “attenuated” by erosion and filling, often with a polyphase evolution (e.g., colluvium effectively felts the “Séré de Rivières” polemoforms of the Nogent-L'Abbesse hill (fig. 1E, 7A-C), as well as the WW1 polemoforms of the Argonne forest (Brenot et al., 2017) without eliminating them.

• “derived” (deep morphological modifications; e.g., Douamont fort in the Verdun belt draped with orange peel modelling during WW1) (De Matos Machado, 2018) (fig. 1D);

• “leveled”, taphonomic structures disappearing from the landscape through erosion or dismantling (Berru battery), or backfilling (Witry-les-Reims fort; trench networks), detectable by the geophysical or archaeological approaches (Devos et al., 2021);

• “regenerated” trenches, generally exhumed by archaeologists or for memorial purposes, such as the trenches network at the Main de Massiges in chalky Champagne (Allouc et al., 2018).

60In addition to detecting and mapping these morphologies, the investigation of their state of preservation is an essential prerequisite for the study, management and promotion of this dark heritage (Hesse, 2014). The particular shape of the battlefields should also be studied in the context of the inscription in the UNESCO World Heritage List of the funeral and memorial sites of the Western Front of the Great War.

61If this contribution testifies once again the need of cross approaches for the understanding of landscapes, it also raises many questions, particularly on the territorial and ecological resilience of these polemoforms, pedoturbated and bombturbated, now islands of biodiversity (LIFE Natur-army). An environmental, pedological, pedo-geochemical and botanical diagnosis would allow a better understanding of these sites, as well as an exhaustive inventory based on the national “LiDAR HD” coverage currently in progress in France (Ancelin et al., accepted a).

*Auteur correspondant : +33 (0)3 26 91 36 83, pierre-yves.ancelin@univ-reims.fr (P.-Y. Ancelin)