Short history of the underground research
For a long time, the study of karst was limited to the surface forms, and these are relatively very poor. Research into the subterranean world of Normandy began very early [Deschesnes and the sinkholes of the Avre Basin, 1675 ; Dezallier d’Argenville, 1755 ; Guettard and the ’Dry Iton’ river, 1758] and continued up to the first world war [Leboullenger and the caves of Caumont, 1801 ; Passy, 1832, 1873 ; Father Cochet and the underground archaeology ; Marchand and the water microbiology, 1854 ; Luard, 1860 ; Fromage and the water resources of the River Robec, 1868 ; Meurdra, 1877 ; Lennier and the Havre region, 1891 ; Ferray and the water resources in Eastern Eure county, 1895 ; Houdry, 1900 ; Fortin and karst infillings, 1906 ; Coutil and the Andelys caves, 1910 ; Dollfus and underground water, 1925-1927], and from the start it included a wide range of science disciplines (geology, hydrology, biology, archaeology, etc.).
However the true birth of the Norman speleology can be attributed to Auguste Monton who in 1894, at the age of 19, completed the first cave survey in Les Andelys county (Eure) [Monton et al., 1991]. He organised the visit of Martel in 1906, and took the first underground photographs. It is for this reason that the caves of Normandy were already well-known when Martel studied them in his Nouveau Traité des Eaux Souterraines , and later in his famous France Ignorée . Curiously, between the two world wars, little new underground research was undertaken, so that it is not until the development of popular caving in the nineteen sixties, that Norman endokarst becomes a focus of interest once more. Because the Lower Normandy caves are rare and small, exploration has been concentrated on the Lower Seine valley.
1- the Norman karst, relation to the regional geology
Normandy is a region in which the crystalline proterozoic Armorican Massif and the sedimentary Paris Basin meet (fig. 1). Since karst is limited to the carbonate rocks exposed to a temperate climate, the greater part of the Armorican region is not of interest here. This region comprises two fundamental geographic areas : the relicts of the Armorican Massif in the west, and the sedimentary plateaux of the Paris Basin in the east.
1.1- the Armorican Massif : testimonies of a relict-covering carbonate rock
Limestones are very rare in the Armorican Massif. However, in spite of this rarity there is a great variety of karst galleries.
Two regions hold research potential.
The first is the littoral plateau of Montmartin sur Mer (Manche), near Coutances, developed in the Visean limestones (Dinantian). A small lime kiln museum is a legacy of quarrying of limestones from ancient times. In these quarries, various small caves were explored, where parts of an ancient karst network were intercepted by the quarrying activity [Druet, 1972]. Unfortunately, urban expansion and the security necessities of the modern society are responsible for the destruction of main parts of this unique karst heritage.
The second main site is located near Saint Lô (Manche), on both banks of the River Vire, in La Meauffe [Juignet, 1961-1962]. At one time, there were quarries opened here in a Brioverian limestone remnant (Proterozoic), that cut through a karst network. Previously once the exploitation had finished uninhibited access was possible with many damages. Today this site is very degraded and is now being infilled by rubbish, even though a study of karst has never been undertaken.
The rarity of karst in the Armorican Massif is reinforced by the lack of studies of a marginal scientific domain, but original and extension of the studies on the Armorican craton. Certainly there have been other karst developments but they have not been mentioned by the very few local investigators.
1.2- The Paris Basin : the domaine of the limestone plateaux
The limestone plateaux of the Paris Basin are better known but this knowledge also reflects the greater interest of the regional research organisation. Two well defined areas can be identified in the region.
1.2.1- the Jurassic ’Campagnes’
As a result of the ancient activity of the University of Caen, research into the natural environment has been well developed in Lower Normandy. However, although a long-standing geological school working for various generations was based here, and despit the intense underground exploitation of building stones under the city, the karst research has been very limited here. Most of the recent discoveries result of the intervention of Upper Normandy researchers (River Aure sinkhole network, Ranville karst, etc.).
Three areas, can be differentiated from south to north, respectively :
the Campagne d’Alençon that includes the famous water karst system of Vingt-Hanaps [Doré et al., 1977]. Developed on the contact of the limestone strata where they ovelie the Armorican Massif, this system is drained by frequent sinkholes. The waters are derived from the Forest of Ecouves, to the springs of the Sourtoir, water-catchment to the Sémallé and Larré communities. The complexity of the karst in this area indicates that an interesting potential study could be undertaken here.
the Campagne d’Argentan about which nothing is known. This may be because there has been no research or possibly for other reason.
the Campagne de Caen-Bessin is the best known and the richest area. It is developed around the River Orne valley (Mondeville, Ranville, etc.), and further to the west, in the Bessin region.
Here the impressive karst system including the sinkholes of the River Aure is developped, extends for over 5 km. Here the visitor must be very careful because of the hydrological sensibility of the system to overflooding from the upper water’s basin, developed in the impermeable layers of the Armorican Massif (region of Caumont l’Eventé).
The system is also sensitive to the flood tide that limits exploration of the downstream part of the cave. These risks result from the very low elevation of the galleries. Because the only modern entrance is near the main sinkhole, exploration can be very dangerous.
Moreover, almost nothing has been published on this fascinating cave system.
1.2.2- The Cretaceous ’chalk lands’
Over recent decades the chalk lands have become the showcase of Normandy speleology. Previously, the chalk was not known to develop karst networks, but this vision has changed over the last thirty years with the publication of works by a few researchers [Rodet, 1992].
In fact, most of the caving and karstic research can be attributed to the activity of Upper Normandy researchers, since the end of the last world war, and more specifically since the nineteen sixties.
This movement is connected with sociological factors such as the establishment of the University of Rouen which was created during the nineteen sixties, and the University of Le Havre in the last twenty years.
The geographical origin of the research can be identified : the Caumont area has been studied by the Rouen caving associations, the Alabaster Coast and the Lower Seine valley (Villequier, Le Landin, Les Andelys, etc.), have been studied by Le Havre cavers, as well as cavers from Elbeuf (Orival, Freneuse, etc.). This aspect demonstrates the sociological behaviour that strongly characterises the regional community of cavers.
Karst regions can be defined by the geomorphological structures that intersect the plateaux, more specifically the seacliffs of the English Channel (Alabaster Coast) and the deep canyon of the Lower Seine valley (fig. 2).
the plateaux of the English Channel margin (Petit Caux or Talou county at east and the Pays de Caux at west),
the plateaux surrounding the Lower Seine valley (Pays de Caux and Norman Vexin on the right bank, Madrie, Roumois and Lieuvin on the left bank),
the plateaux of Southern Eure (Plateaux de l’Eure-Saint André, Pays d’Ouche and Plateau du Neubourg),
the hills of the Perche (Orne county).
The two first sectors are the birthplace of physical speleology in Normandy. The plateaux of the Southern Eure hold great potential for discovery because there are many impressive karst connections between river sinkholes and springs. However, the research stays too much difficult and limited. In spite of this the hills of the southern district of the Perche have shown, with considerable work by the CNEK during the nineteen ninties, an exceptional potential that has been weakly exploited because of its great distance from the larger cities and because of the rural way of life in this area.
2- The chalk karst, a Norman speciality ?
The first quality of the Norman karst is its development, for a large part, in the porous limestones of the Upper Cretaceous chalk (Cenomanian, Turonian, Senonian). The consequence is that normally the porosity, responsible for the dispersion of water into the carbonate structure, does not allow the development of karst, or the concentration of water. Despite a porosity over 40 % of the rock mass, karst is developed. The evidence is clear from the abundance of caves explored, surveyed, photographed and studied.
For a long time, this evidence was denied by many researchers, some of whom had an international reputation, but the evidence could be seen by thousands of tourists : including the impressive Trou à l’Homme cave dug into the well-known cape of the "Porte d’Aval", one of the natural arches at Etretat.
2.1- Chalk and karstification : a Norman model ?
The chalky limestone strata of the Normandy plateaux present a large variety of facies but the hydrological characteristics are very similar to one another because of the major role played by the porosity in the karst. The low regional elevation (under 300 m a.s.l.) and the lack of main tectonic features (with the exception of the eroded Bray anticlinale) limit this geomorphological expression beneath continuous weathering covering of several metres deep. This is sometimes reinforced by elements of pre-existing Tertiary strata that conceal the chalk beneath. In all this region, chalk is only visible in the sea cliffs of the English Channel, northwest of the region, and in the deep valley of the River Seine that cuts through Eastern Normandy, from SE to NW. These natural geological sections reveal a large number of karst phenomena, more specifically fossil entrances and springs. The input karst landforms (potholes, pipes, sinkholes, etc.) develop under the weathering cover and constitute the active front (cryptokarst) but they are not normally being visible.
Moreover, the development of many quarries in the overlaying beds or in the chalk bedrock make access more difficult.
Chalk is a limestone the porosity of which allows the development of a watertable. This aquifer is drained by hydrogeological linkages resulting in karstification of the limestone bedrock. The weak water gradient does not have direct impact on the drainage of this dual aquifer which develops in all carbonate areas, at various stages of evolution. A continental area can be identified where karstic networks represent long stability phases of the drainage conditions (terrestrial processes), and a coastal area can be identified where Quaternary sea-level variations have lead to destabilisation and breaks in the drainage conditions. It is under these ultimate influences that water could penetrate into a large area in the region following lower valleys like the River Seine, that exacerbated endokarstic organisations comparable to the development of a delta.
Spectacular underground piracies of epigenous rivers by karstic systems can also be seen developed from the sea coast (Veules, Vaurain, Yport, Bébec). Cryptokarst, hydrogeological links and underground collectors with exogenic infills are representative of a lower plateau karst system. The sea coastal influence is responsible for the exacerbated development of the output karst sector and the underground drainage that catches waters from valleys and dales.
2.2- The ’primokarst’ of the Perche country
The Perche county is a region famous for its varied landscape with hills covered by meadows and deep forests, opened by valleys with many rivers. Between the capital Mortagne-au-Perche and the main town Nogent-le-Rotrou, the upper valley of the River Huisne, tributary of the River Sarthe, extends into a large basin dominated in the north by the crest of the Perche cuesta. In this depression, some ridges appear, hollowed by rivers, in which a countless multitude of underground quarries where dug for agriculture or for building. These underground chalk quarries, in lower and middle Cenomanian limestones, have cut through surprising natural voids, completely infilled with clay.
Systematically, these voids are directed by vertical and subvertical joints and sometimes by the bedding (fig. 3). Voids without a leading joint do not exist.
There are weathering fronts located on the joint network, that selectively present a spectacular widening at the contact with an old watertable level.
These weathering fronts still function today but it is not demonstrated whether they are functioning as much as in the recent past.
Consequently the studied caves are not galleries enlarged by concentrated floods, but they are mainly vertical solution voids, opened in the bedrock by input water percolation.
This is why each explored part has to be excaved from the residual complex or ’rock-ghost’.
This was achieved mainly by anthropogenic action, sometimes by rupture of the insoluble residuum, evolving from an iso-weathered stage to an allo-weathered stage, with aeration and drying and packing down of the ’rock-ghost’.
In this last stage, the weathering front can trap insoluble elements from the upper part of the bedrock and sometimes from the surface.
If sufficient water begins to flood in the joint axis, that may organise a karstic drainage ; in this case it is possible to speak of an initial karst stage or ’primokarst’ [Rodet, 1996, 1999].
2.3- The chalky ’causses’ of the southern Eure country
Downstream in the Perche county, a regional water supply (the spring of many rivers : Sarthe, Orne, Touques, Guiel, Charentonne, Risle, Iton, Avre, Eure, Loir, etc.) established over the Perche Sands Formation (Upper Cenomanian), develop areas where these rivers lose part or all their water into sinkholes.
This mechanism is amplified by lateral facies change from a sandy-clay facies (Sables du Perche) to the Rouen Chalk (Upper Cenomanian white sandy facies).
The physical-chemical dye tracing experiments have shown karstic links with impressive springs, sometimes very far downstream from the sinkhole (up to 27 km in a straight line), with high minimum average speed (over 200 m/h).
Here an unpenetrable cave is known today because the sinkhole is infilled by insoluble sediments. However, this does not mean that galleries do not exist in the area.
2.4-Relict karst of the Lower Seine
The Lower Seine valley offers a large number of karst sites, with different geometry and morphology.
Nevertheless, each one presents a same aspect : none is in accordance with its geomorphological context. In fact they are developments inherited from a different context. We can classify these karsts into two or three altitudinal groups on the valley sides.
The first group occurs at the top of the hill slope, i.e. various tens metres above the river level (Les Andelys, Freneuse, Orival, Moulineaux, Duclair, etc.).
The second group occupies a much lower position, ten to twenty metres above the river (Canteleu, Le Landin, Saint Arnoult, Caumont, etc.).
A last group occurs at the level of the modern valley floor (Gouy, Villequier, Radicatel, etc.). Those two last groups are, more or less, the same entity linked with the evolution of the great valley during the Late Pleistocene (estuary and river environments). By contrast, an important gap between the higher group and the second one is evident. In this first group most of the dismembered galleries are most difficult to place into their context.
This break might result from the different processes activing between these two significant karstic phases : between both, or this intervening phase was never karstified, or the evidence was destroyed by valley resizing that spared the topmost karst networks. This resizing could result of the introduction of the lower valley into the estuarine domain or/and of the capture of the Marne Basin, an ancient tributary of the River Meuse. In this way, two kinds of paleokarst can be distinguished :
the upper group, from Lower to Middle Pleistocene age, was deconnected by the river valley deepering and dismantled by the valley-side evolution. Here the galleries rather present a simple morphology, normally, with limited stepping.
the lower group was broadly developed under the influence of the altitude variations of the Upper Pleistocene, and should be considered as a littoral karst, with its complex organization with stepping and superimposed galleries. In such an evolving context Holocene karstification did not really occur and thus become evident through morphological retouches to older karst networks.
3- Karst and the coast
Normandy is a particular region in which the impressive spatial development of the karstic forms and system organisations formed under the influence of the Quaternary sea-level changes, can be seen.
3.1- Coastal development ; a Quaternary history
The Petites Dales cave (fig. 4), 1,3 km from the sea shoreline, is a major cave in Upper Normandy. Its 550 m of galleries represent specifically a large collector (2-3 m wide, 10 m high), that was completely infilled before it was opened by cavers. Its conduit size, the absence of stepping in the galleries, and the absence of a conduit confluence or diffluence in the upper part of the collector, and the absence of a by-pass gallery illustrating a destabilisation or a significant migration of the drainage axis, and the important elevation of the per ascensum digging ; together these features, all elements of terrestrial evolution, indicate a long stable period of water level and of the drainage conditions. Despite the proximity of this cave to the coastline border it was never influenced by littoral processes. It seems that drainage conditions were only modified by the Petites Dalles valley incision that alloved the water to flood away from the aquifer. This valley incision was only deepened as the coastline approached. Therefore it seems that the cave drainage was fossilised by the dale incising under the karst network base.
Caves developed under such conditions, are frequent in the Alabaster Coast region, demonstrating that they are not exceptional geomorphologically. In fact, it is possible to distinguish the ’littoral’, a geomorphological zone where the sea and the continent confront each other, and the ’littorallity’, a temporal concept of a conflict between marine and continental processes, that migrates in relatively space with the temporal evolution of the relative mean sea-level during the Quaternary. In this way, a continental cave can be seen on the coastline never retouched by the sea. It can also be seen far away from the modern coast, in the side of a deeply incised valley like the Lower Seine. There a karst network wholly subjected to Quaternary fluctuations of the sea-level. This is why it is very important to distinguish the present coastal zone from that influenced by previous Quaternary sea-level changes.
3.2- Seacliffs and karst
The Alabaster Coast zone is a specific region very similar to that of the Lower Seine. Nevertheless, it owes its form to the seacliff rempart that recedes during transgressions and fossilises under continental deposits during regressions and sea-level lowstands (Senneville sur Fécamp). Therefore palaeogalleries can be frequently observed perched at the top of seacliffs (Bruneval), whereas at the wallfoot ascending springs develop (Senneville sur Fécamp). Some drainage, established over impermeable horizons, remain hanging above the sea-level by the recession of the shore cliff (Antifer’s Cape) and its waters sometimes precipitate calcareous tufa (Bénouville).
Consequently, palaeodrainage, including in coastal erosion zones, are responsible for the development of spectacular landforms such as the Etretat arcs.
The chalk seacliffs result to a greater extent from continental processes rather than from marine abrasion. The marine abrasion is responsible for the mass evacuation of the unprotected active cliff-foot, and the verticality of the cliff results from the evacuation of collapse products. They also result from terrestrial, often karst waters, that infiltrates the bedrock and exploit the tension fissures, that are more specifically attacked by frost action. The coastline of the carbonate area of Normandy is under direct karst influence. During the Quaternary, three evolutionary stages can be distinguished :
1- the coastline is distant from the area concerned : the continental hydrological basins are extended and stable, drained by substantial karstic systems (Petites Dales cave),
2- the coastline approaches the area concerned : the extended continental basins are divided into littoral basins where dales incise.
3- the coastline reaches the area concerned : dales are fossilised and become disconnected. New karstic drainage develops under the plateau, without any evident relation to the landscape (karstic capture of hydrographical basins). The first stage totally fossilised palaeokarsts are exhumed by the shore cliff retreat.
3.3- Underground captures
One of the consequences of the transgressive periods is the modification of karstic processes into the chalky massif. These processes radically overturn the local drainage conditions, by modern (Yport) or ancient (Jambourg Dale at Etretat) capture mechanisms. These karstic axes predetermine the ways in which carving and the intense erosion of the sea cliffs by the coastal processes occur. The ’Pertes de l’Aure’ system (Calvados) seems to offer a better example of this process. This process does not only affect the chalk limestones, but also the Bajocian limestones of the Bessin district. According to previous researchers, the River Aure first flooded from the south (on the basal complex near Caumont l’Eventé) to the north as far as the town of Port en Bessin, passing by the ’Plaine de Port’ plateau, through the cuesta front sedimentary strata. In a second period, the Early Pleistocene tra nsgression of the ’Col du Contentin’ bay drove the development of the drainage to the western Veys Bay. Incision of the valley was aided by an old east-west fault that forced the river to change its course 90° to the west. The retreat of the shoreline in the Bessin district during transgressions, favoured by a more efficient hydraulic gradient brought about the karstic capture of the waters of the River Aure, near the bend where Bajocian layers outcropt. From then on the valley is drained by two different rivers. The Upper Aure, from Caumont l’Eventé, through Bayeux, disappears into many sinkholes, one of which, the ’Fosse Souci’ is the most spectacular, with a 9 m3/s flood input. It discharges on the shoreline, around Port en Bessin. Close to the last sinkhole, on the other side of a small pass, the Lower Aure river draines the impervious beds overlying the limestones, and flows towards the Veys Bay. During floods, all the sinkholes saturate and the Aure overflows into the lower valley. The different physico-chemical quality between the Upper Aure and the Lower Aure demonstrates that two different rivers drain the same valley.
3.4- Karst as a signal of the Quaternary shoreline evolution in the lower Seine valley
The plateaux of the Lower Seine limit the depth of incision of this great valley, between the confluence of the River Epte and the estuary in the English Channel. The Quaternary evolution is complex and recorded in the karst development. The upper part of the valley always was a river (terrestrial environment) when the downstream part, more or less important, was included in the estuary, or even marine during transgression, or a river environment during regression. This was dependent on global sea-level fluctuations. These changes resulted in large modifications of the hydrological base-level and in consequence of the karstic drainage. This study contributes to the chronological reconstruction of the Quaternary evolution in the Lower Seine.
In this way, the Caumont area, 120 km distant from the coastal zone along the River Seine, at an elevation of only 4 m, is under the direct influence of the tidal zone (fig. 6). The large underground network, known to exceed 7 km, includes a number of different conduits, with several connections, developed in a complex organisation scheme. This results from the adaptation of underground drainage to the various variations of the hydrological base-level rather than from the geomorphological variations of its hydrological basin. In fact, the considerable distance of this network from the sea-cliff erosion, protected the Caumont underground system, with its rich evidence of complex Quaternary evolution. This explains why the modern coastal zone, subjected to the erosional effect of a rejuvenated shore-line, does not provide such impressive karst examples (Cap Fagnet) like those at Caumont, which is protected by its upper estuarian location.
The chalk karst in Normandy can be characterised as follows :
1- the development of a dual aquifer, composed of karstic drainage responsible for the quality of waters, and of a phreatic water-table that results from the porous carbonate bedrock giving the quantity of the water resource.
2- the developement of a very dynamic but poorly known cryptokarst, under a thick weathering layer covered by exogenous deposits (loess, sand, etc.).
3- morphological evidence of different evolution steps of the karst, more specifically the ’primokarst’ with its ’rock-ghost’.
4- a low plateau position that prevents strong hydraulic gradients. Quality and quantity of the underground infillings demonstrate the operation of these weak processes.
5- a littoral location during the Quaternary, involving exacerbated development and the organisation of the output area, the ’karstic delta’, and also the fossilisation of underground networks that evolved into palaeokarsts.
Doré F., Juignet P., Larsonneur C., Pareyn C., Rioult M., 1977. Normandie. Guides Géologiques Régionaux, Masson, 207 p.
Druet Y., 1972. Approche des phénomènes karstiques en Basse Normandie au travers de deux départements de référence : la Manche et le Calvados. Mémoire de Maîtrise, Géographie, Université de Caen, 125 p.
Evrard H., 1987. Risques liés aux carrières souterraines abandonnées de Normandie. Bulletin de Liaison des Laboratoires des Ponts et Chaussées, 150-151 : 96-108.
Juignet P., 1961-1962. Géologie de la région de Pont-Hébert – Airel (Manche) : le Briovérien et les calcaires de La Meauffe. Mémoires de la Société Nationale des Sciences Naturelles et Mathématiques de Cherbourg, Tome L, 5ème série, 10 : 121-199.
Monton J., Rodet J., Viard J.-P., 1991. Auguste Monton, spéléologue des Andelys. Actes du Museum de Rouen, 1991 (2) : 17-28.
Rodet J., 1992. La craie et ses karsts. Ed. CNEK-Groupe Seine, Elbeuf, 560 p.
Rodet J., 1996. Une nouvelle organisation géométrique du drainage karstique des craies : le labyrinthe d’altération, l’exemple de la grotte de la Mansonnière (Bellou sur Huisne, Orne, France). C.R. Acad. Sci. Paris, t. 322, série IIa : 1039-1045.
Rodet J., 1999. Le réseau de fracturation, facteur initial de la karstification des craies dans les collines du ¨Perche : l’exemple du site de la Mansonnière (Bellou sur Huisne, Orne, France). Geodinamica Acta, 12 (1-2) : 259-265.
Rodet J., 2003. La Normandie souterraine. In La Normandie, A. Guérin dir., Bibliothèque du Naturaliste, Delachaux et Niestlé, 52-63.
Rodet J., Masséi N., Laignel B., Dupont J.-P., 2003. The karstic delta as a morphological consequence of base level variations. Example of a chalk karst system in the western Paris Basin (Normandie, France). Climate Changes : the karst record III, Montpellier, 11-14 may 2003, poster : 139-140.