Précédent Suivant

Chapter 4 – Hydrogeology, Infiltration Water Characteristics and Support Properties at Lascaux

p. 81-89


Texte intégral

4.1 Introduction

1Lascaux Cave was discovered in 1940 near Montignac in Dordogne, on a hill above the Vézère Valley. Due to its shallow position and intrinsic properties—low volume, very high CO2 levels—this treasure of humanity quickly proved to be very fragile. This state of affairs caused the authorities to close the cave in 1963 and since then to study the functioning of the cave.

2Many studies have been conducted on the prehistoric art and methods for its preservation. The first geological and hydrogeological surveys carried out by the University of Bordeaux after the cave was closed (Schoeller 1965; Vouvé 1968) clarified its geological and geomorphological environment. Many questions remained unanswered, however, with regard to the hydrogeological and climatological workings of the cave due to the lack of methodological and scientific tools available at the time. Similarly, knowledge of karst environments was in its infancy and it was not until the work done by French teams from the University of Montpellier, the Centre National de la Recherche Scientifique (CNRS) at Moulis in Ariège, and the University of Avignon that its functioning could be better understood (Drogue 1969; Mangin 1975; Bakalowicz 1979; Williams 1982; Mudry 1987; Lastennet 1994).

3Following the microbiological attack of 2001, the recently appointed scientific committee launched new research projects, including, following the advice of an expert on the committee, Michel Bakalowicz, hydrogeological studies: while it is urgent to protect the ornamentation inside the cave, it is through an overall understanding of the system that can be achieved. The cave interacts with the external environment and the resulting stresses must be identified and quantified. These new studies, conducted by the University of Bordeaux, funded by the DRAC (Regional Department of Cultural Affairs) and supported by the Region of Aquitaine through their award of a thesis grant1, began in autumn 2003. The summary below is largely taken from the thesis by B. Lopez, which is currently being finalized. The methodology used consisted of monitoring the development of hydrogeological and geochemical parameters, as well as many geological surveys in the field, the results of which are presented below. In parallel with this approach, and following the fungal attacks on the walls, the scientific committee instigated research on another scale, linking the interaction between the microclimate and the development of fungus to the centimeter scale and the pore scale. This research is the subject of a program that is currently in progress, conducted by the Istituto di Scienze dell’Atmosfera e del Clima (attached to the High Council of Research) in Padua, the Institut National de la Recherche Agronomique (INRA) and the Laboratoire de Recherche des Monuments Historiques (LRMH). Some aspects relating to cave support are presented below.

4.2 The physical properties of the site of Lascaux

4.2.1 Geographical context

4Lascaux Cave is located in the Black Perigord area, in the department of Dordogne, near the town of Montignac (fig. 24). The Black Perigord is located on the northeast edge of the Aquitaine Basin. This particular geographic zone is bounded to the north by the A89 motorway (formerly the N89), to the south by the Dordogne River, to the west by the Barade Forest and to the east by an imaginary line from Brive to Gourdon. The Vezere and the Dordogne are the two main rivers that cross this area. The typical landscape of Black Perigord consists of plateaus that often peak at an elevation of 300 to 350 m NGF and valleys at altitudes of close to 100 m NGF. The hydrographic network is dichotomized and highly differentiated into small sub‑basins.

Image 100000000000046500000330C4008C3EE6F1A3C3.jpg

FIG. 24 – Location of Black Périgord region and Lascaux Cave.

© Alain Dagand.

5The climate of the Dordogne is mainly oceanic, although it does have some continental influences, as shown by occasional snowfall in winter months. There is an average precipitation of approximately 880 mm per year, split into two main periods: autumn and spring. The average ambient temperature at Lascaux is approximately 12.6°C, with monthly temperatures ranging from 3°C in winter to 20°C in summer. The hill on which Lascaux is located is a butte of the Upper Cretaceous limestone plateau of the region. It is flanked on the north and south by two valleys with streams running through them, and bordered to the west by the Vezere valley (fig. 25).

Image 100000000000063F000006A2B7B5CEE8D2142B62.jpg

FIG. 25 – Geographic context of Lascaux Cave and photos of remarkable sites.

Extrait de la carte topographique n° 2035O : © IGN ‑ Paris 2010, autorisation 80‑0063. Clichés : © B. Lopez / université Bordeaux 1, laboratoire Ghymac.

6The geological history of Black Perigord is characterized by abundant Jurassic limestone sedimentation. Its emergence and peneplanation occurred starting in the Lower Cretaceous period, until the sea again covered the region during the Upper Cretaceous. The sediments deposited are thus of two types:
– detritic (fossilized sandstone with mostly spathic calcite cement) during the Turonian, Coniacian and Maastrichtian periods;
– terrigenous during the Santonian and Campanian periods.

7During Tertiary period, the Siderolithic formations and Perigord sand formed a coating over the preexisting reliefs. This covering is made up of sand and gravel distributed in an argillaceous paste that varies from brown to reddish.

4.2.2 Geological and geomorphological context of the site of Lascaux

4.2.2.1 Geological map (after Schoeller 1965)

8The stratigraphic sequence of the hill is as follows (figs 26‑27):
C6c – Ferruginous sandstone and calcariferous sandstone of the Upper Turonian period (10‑25 m thick). These outcrop along the small cliff of Puy Robert, and were found at the source of the Fageotte stream during channeling work.
C7a – Marls and marly glauconitic limestones from the base of the Coniacian, at Barroisiceras haberfellneri, etc. (approx. 10 m thick). They outcrop under Lascaux Castle, at the top of the Puy Robert cliff, along the road to the cave above La Fageotte, above the Regourdou road and to the east of La Grande Béchade.
C7b – Coniacian limestone, yellow limestone forming the structure of the hill (100 m thick). C7c – Arenacious Santonian limestone, constituting the Régourdou plateau.
SP – Fine grained, yellow sands, often with an argillaceous stratum at the base. They cover the hill that dominates to the east of Lascaux Cave.
a1 – Ancient alluvial deposits between Lascaux Castle and Puy Robert, which also form the Grande Guionie terrace.
a2 – Recent alluvial deposits of the Vezere River.
Aa – Alluvial fans to the east of La Béchade and La Fageotte.

Image 10000000000005690000048BD44B676F6972F38D.jpg

FIG. 26 – Geological map of the Lascaux hill (after Schoeller 1965).

© B. Lopez / université Bordeaux 1, laboratoire Ghymac.

Image 10000000000002E1000003FF7FE787A2EDD1FE12.jpg

FIG. 27 – Lithostratigraphic log of the geological formations of the Lascaux hill.

© B. Lopez / université Bordeaux 1, laboratoire Ghymac.

9The cave developed through karstic erosion in the upper part of the Coniacian limestone. This limestone, which is well bedded and jointed on the surface but very compact farther down, is visible just above the cave. On both sides of this promontory, the limestone is covered by a sandy, argillaceous formation that extends beneath chestnut tree cover on the top of the hill. This formation includes “sables fauves” (mixed fine and coarse grains), that are argillaceous, with traces of slightly ferruginous mica impregnated with manganese oxide along the veinules. Light gray or ochre clay with traces of mica can be found within. The iron oxide impregnations tend to have a reddish tint. This clay directly covers the limestone.

4.2.2.2 Geomorphological and structural approach

10In 1968, Jean Vouvé succeeded in defining the geomorphology of the zone near Lascaux Cave. Through the use of direct auger surveys he was able to identify very deep furrows filled with clayey‑sands.

11Between 2005 and 2008, we defined this geomorphology more precisely by combining non‑destructive geophysical methods and geotechnical methods consisting of penetrometic surveys outside the wall that protects the cave. Similarly, a structural measurement survey was conducted at the scale of the hill to detect the fracturation and karstification directions.

12The carbonated base that composes the structure of the Lascaux hill is widely jointed and fractured. Based on a methodical survey of the fracturing planes in situ, four sub‑sets were defined:
– F1, of average direction N178 (s = 11.3), 19 readings;
– F2, of average direction N120 (s = 4.9), 18 readings;
– F3, of average direction N93 (s = 6.8), 5 readings;
– F4, of average direction N145 (s = 7.2), 10 readings.

13Most of the fracture planes are sub‑vertical, with one exception found in F1 (number of readings = 19), three in F2 (n = 18) and one in F4 (n = 10). Tracing the average S0 stratification and fracture planes in a Wulf diagram (fig. 28) allows us to summarize the different structural sets identified on the site. This fracture reading (fig. 29) partly explains the karstogenesis of the cave channels. In effect, the Right Gallery and the Mondmilch Gallery were developed in relation to the F1 fracturing of average direction N178. The Hall of the Bulls and the Axial Gallery were hollowed out along a N115 fracture plane belonging to set F2, and which is intersected by F1. Likewise, the main direction of the Shaft (N85) corresponds to the F3 fracture system, which also cut the limestone promontory into two furrows in the northwest of the cave. Finally, even if set F4, in the Armorican direction, does not support a karst channel belonging to the site under study, it is nevertheless represented at the scale of the hill by deep thalwegs, as well as by the filled paleocanyons on both sides of the cave.

Image 1000000000000333000001B93A37317CA72DBC2A.jpg

FIG. 28 – Representation on a Wulf diagram (Southern hemisphere projection) of mean S0 (stratification map) and fracturation levels measured on the Lascaux hill (Middle and Upper Coniacian, and Santonian).

© B. Lopez / université Bordeaux 1, laboratoire Ghymac.

14The natural site of the hill constitutes a broadly homogenous geomorphologic ensemble. It is mostly covered by a mixed forest where evergreens (Pinus sylvestris) alternate with deciduous trees (chestnuts and oaks). Beneath the forest cover, the soil is continuous and varies in thickness depending on the nature of the substrate. It is thickest in the cultivated clearings, while in the uncultivated clearings, or where there are grasses and junipers, there is no soil horizon and the rock is bare. The geological foundations that delimit the hill constitute an ensemble of superimposed strata that are horizontal and varied. In the upper part, the many excavations carried out starting around twenty years ago are clearly visible; they descended into the calcareous sandstone that is altered, fissured and karstified in some places and which is attributed to the Santonian period. The unaltered rock has yellowish‑brown tints, the altered passages have been hardened by water flows and the rock matrix is mainly orange‑brown or rust‑colored. A long series of detritic and biodetritic limestones attributed to the Coniacian period formed below this. The thickness of this formation attains around 80 m. The upper fossil and ornamented system of Lascaux is situated within this limestone, and is extended downwards by an unknown lower system. This level is limited at its base by a marly, gray‑blue facies visible on the excavated side of the access route to the cave.

15To improve our knowledge of the geometry of the formations near the cave, geophysical and geotechnical surveys were conducted. Geophysical methods are particularly suitable for conservation studies because they are not destructive. Two electromagnetic survey sessions (EM 31) were thus carried out, one in April 2006 in the fill formations to the northeast, and the other, in April 2007, within the enclosure protecting Lascaux Cave (Peyraube et al. 2008). The moisture conditions of the surface and subsurface during these two campaigns being comparable, we combined the two isoresistivity charts measured at EM 31 to provide an overall image of this surface characteristic over the whole site. Two maps were thus drawn up: the resistivity appearing at a depth of approximately 2 m, measured by the horizontal field (fig. 30) and the resistivity appearing at a depth of approximately 5 m, measured by the vertical field.

Image 10000000000005F10000055BCC824B10B55CC257.jpg

FIG. 30 – EM 31 isoresistivity map: horizontal field, investigation depth: approx. 2 m (top); vertical field, investigation depth: approx. 5 m (bottom).

© B. Lopez / université Bordeaux 1, laboratoire Ghymac

16The main information obtained from these surveys is the clear heterogeneity of the observable resistivity at the site, which can be explained by the very particular geomorphology of the epikarst system at Lascaux. The cave is situated in a limestone promontory (high resistivity value) that is severely eroded into paleocanyons and filled with sandy, clay material. The range of resistivity variations indicates significant variability in the lithology of the fill, which has a higher clay content when resistivity is low and more sand when the resistivity is higher. Through these studies, the geometry of these formations is now fairly well known, at least under the surface. To obtain more information, a dynamic penetrometer campaign was undertaken outside the enclosure to identify the dimensions of the limestone substratum. This prospection has confirmed the significant lithological variability and the particular geomorphology of the limestone. The canyons identified can reach a depth of 15 m, with the substratum coming close to the soil in some places.

4.3 Hydrogeology

4.3.1 General hydrogeological context

17Lascaux Cave is an ancient karstic channel that developed in the Upper Coniacian formations. It is part of a karst massif whose extension is that of Lascaux hill, but whose outlets take the form of several slow flowing rising springs (Haute Fageotte springs, La Madeleine spring: fig. 25), or an even more diffuse form traversing the Lower Coniacian formation, though it is thought to be less permeable. We indeed observe that the valley’s major springs come from deeper strata since they are warmer than the local thermal gradient. There is therefore no single outlet that drains the massif of Lascaux hill. The studies carried out on the site indicate that the entire cave, the limestone promontory and the paleocanyons can be assimilated into an epikarst system whose main outlet is the exsurgence situated in Lascaux Cave, in airlock chamber 1, compartment 3 (fig. 31). This is not the only outlet, since other water exits are occasionally visible in the cave (Chamber of the Felines, upper part of the Mondmilch Gallery, inter‑ledge joints). The water in airlock chamber 1 is collected on runners (fig. 32) and then channeled to a tank buried in the machine room. It is then raised toward the exterior by means of a submerged pump, which allows the groundwater flow to be measured.

Image 10000000000004800000027F73457EC055CD26B9.jpg

FIG. 30 – EM 31 isoresistivity map: horizontal field, investigation depth: approx. 2 m  (top); vertical field, investigation depth: approx. 5 m (bottom).

© B. Lopez / université Bordeaux 1, laboratoire Ghymac

Image 10000000000002EC0000021302B8591CED397ED8.jpg

FIG. 32 – Outlet zone on the ceiling of airlock 1, section 3, of Lascaux Cave, and water collection system.

© B. Lopez / université Bordeaux 1, laboratoire Ghymac.

4.3.2 Functioning of the epikarst system

4.3.2.1 Hydrodynamic approach

18A particularity of the flows at the Lascaux Cave is their non perennial nature. In fig. 33, we can see that the flows generally begin in late autumn and finish at the end of spring, in June. In exceptional cases, flows are recorded during the summer season—in 1997, for example. A study of the relationship between rain and flows, using the approach of Alain Mangin to signal processing (Mangin 1975) on flow columns on a daily basis since 1965 and every 10 minutes since 1997, shows a fairly rapid reaction by the epikarst system, with a flow response time of one to two days on average, depending on the cycles, after rainfall. The hydrograph is very typical of a karst response by the Lascaux system, and not of an environment with interstitial porosity (Lastennet et al. 1999). There is significant variability, however, due to the soil and epikarst refill state, which allows or refuses pressure to be displaced to the outlet. Studying the hydrograph clearly shows that the flows are subtended by the refilling of the system in autumn. A deficient epikarst refill state leads to a time lag in the flow uptake, which can even be strongly influenced by the deficit caused by preceding cycles.

19These findings are important for the future management of the cave, since we are well aware that, based on analysis of the data, variations in the climate over several years will have a major impact on the overall humidity of the massif, including Lascaux Cave. Meanwhile, these microclimatological studies conducted inside the cave demonstrate the very significant influence of the groundwater on the thermal and thus aeraulic functions of the cave (cf. chapter 5).

20Processing the rain/flow signal reveals another basic parameter required for the characterization of a karst system, namely the regulation time, which takes account of the storage capacity and drainage of the reservoir that feeds the outlet. This consists of knowing whether rainfall or a rainy period will have a lengthy effect on the flows recorded at the outlet. In the case of Lascaux, there are two types of functioning: some signals show a regulation time of 11 to 12 days, while others show a time of around thirty days. According to our current hypotheses, these different figures correspond to the drainage of separate sections of the epikarst, depending on their permeability and their relationship with the preferential transmitting drainage zone that channels the water to the airlock chamber 1 outlet.

21The sectors that drain with the greatest difficulty correspond to the capacitive function of the epikarst. In Lascaux cave, it is not yet clear whether the karst fill is involved in this function by slowing the flow and allowing a reservoir to be built up. These flows are still probably minor, due to the low permeability of the environment, but they do play a considerable role in the geochemical signature of the waters at Lascaux.

22Some of the more transmissive sectors, composed of fractured limestone, are refilled more quickly and explain the rapid ascent of the flows. On the other hand, they have only a small reservoir and dry up quickly.

4.3.2.2 Mass transfer in the system

23The vulnerability of a karst system is most often assessed according to the speed of water flow within. This usually means taking account of the risk of pollution reaching the outlet. To ensure the preservation of a cave, this risk must also be taken into consideration since caves are very sensitive to what occurs in their immediate environment. Beyond these considerations of potential contamination, meanwhile, rapid flows themselves constitute a potential risk for the walls due to the aggressive nature of fast‑flowing waters in chemical imbalance with the surrounding stone.

24Traditionally, the natural water flow is traced in order to profile the mass transfer (or transit) in a hydrological system (Lastennet et al. 1996; Lastennet, Mudry 1997). Two cycles have been monitored in stable oxygen (18O) and hydrogen (deuterium 2H) isotopes, at the outlet and in the precipitation at the site of Lascaux itself. The resulting curves show a spectacular absorption of the isotopic signal from rain (fig. 34) that has transited the system. The soil and the altered part of the limestone (crumbled limestone) known as “epikarst” filter the “rain” signal by homogenizing the water. The small range of oxygen‑18 variation in the water at the spring indicates that overall mass transfers are slow in these geological formations, since the “rain” signal shows great annual variation. We have confirmation of this in the absorbed signal obtained at the Haute Fageotte spring, situated below Lascaux Cave at the foot of the hill. By establishing the weighted averages of the isotopic signals in the rain (average ‑7,6 δ‰) and at the airlock chamber 1 exsurgence (average ‑7 δ‰), we observe a significant difference, contrary to what is usually found. In fact, the “spring” signal is traditionally more negative than the “rain” signal, since it is the poorer winter rain that tends to refill the hydrogeological systems. This finding proves that the flowing waters result from a mixture of water with those from previous cycles and that this water is pushed out when the epikarst system is filled up.

Image 1000000000000569000003C3FE99C4221156672C.jpg

FIG. 34 – Trend in the oxygen-18 signal in precipitation at the airlock 1 outlet and at the Great Fageotte spring during two cycles, from 2003 to 2005.

© B. Lopez / université Bordeaux 1, laboratoire Ghymac.

25Through more detailed analyses of the isotopic response, by extending the scales, we find that there is a correlation between the fluctuations in the isotopic flow signals and the isotopic source signal. Figure 35 shows the same trends between these signals, demonstrating that over the analysis period (15 days on average) fresh rain has a slight influence on the outlet. This finding has major implications for preservation since it proves that recent water can enter the cave, albeit in small quantities, within a period of fifteen days when the system is active (measurable flows at the outlet).

Image 100000000000063C0000039CF383DD62EF797BF3.jpg

FIG. 35 – Comparison between isotopic oxygen-18 flows in rainfall and the oxygen-18 signal of the airlock 1 outlet.

© R. Lastennet et B. Lopez / université Bordeaux 1, laboratoire Ghymac.

4.3.3 Geochemical characterization of the waters

4.3.3.1 Calco‑carbonic balance of the waters in airlock chamber 1

26The calco‑carbonic balance of the water in the karst system is largely governed by the CO2 contents encountered by the water throughout its course of infiltration. The dissolution capacity of water depends on its concentration of dissolved CO2, which is necessary to attack calcite. The water is quickly balanced with the gas phase, and with every change in the atmosphere in which it circulates. During three cycles (from 2003 to 2006: fig. 36), we have demonstrated that the partial CO2 pressure encountered by the water flowing out of airlock chamber 1 and during their circulation through the system is very high (4.8% on average for cycles 03‑04 and 04‑05). This highly mineralized water is systematically supersaturated with calcite and is therefore deposits incrustations when it flows through the cave. This is essentially what we find on the ceiling of airlock chamber 1, which has many concretions.

Image 10000000000005680000038529BA49D5218D3D91.jpg

FIG. 36 – Trend in partial pressure (p) of CO2 calculated in the water at airlock 1, pCO2 calculated to be balanced with the calcite in the water (pCO2 ISc=0): precipitation at Lascaux and flows at airlock 1 (top), and water saturation value in relation to the calcite (ISc) (bottom).

© B. Lopez / université Bordeaux 1, laboratoire Ghymac.

27This phenomenon is explained by early degassing in the massif, rebalancing the water with an atmosphere containing less CO2. The difference between the CO2 pressure in the air and in the water determines the precipitation mechanism. An important question remains unanswered at Lascaux concerning the pumping of CO2 at the bottom of the cave: what are the advantages and disadvantages for the cave walls of pumping CO2? If the CO2 levels are too high, condensation water could become charged with CO2 and aggressively alter the walls by a process of dissolution. On the other hand, if the CO2 level is very low, the infiltration water that is rich in calcite and CO2 can degas and precipitate, covering the walls with calcite. The growth of calcite crystals in the gaps and fissures below the calcitic veil on the surface might explain the flaking of the wall in some sections.

4.3.3.2 Origin of the CO2: information on the hydrogeological functioning

28The calculated CO2 pressures in the water at Lascaux are much higher than the concentrations traditionally found in karst environments (between 1 and 2%). This raises questions about the origin of the CO2 and the mechanism of such a concentration, as well the impact of the atmospheric pressure on its dynamics (Bourges et al. 2001; Denis et al. 2005; Spötl et al. 2005). Beginning in 1963, Henri Schoeller measured the very high partial CO2 pressure in the water at Lascaux without knowing its origin. Some abnormal CO2 pressure was also measured in the air of the cave itself, in the Shaft (up to 8%). In 1975, Jean Vouvé identified the CO2 abnormalities in the detritic filling close to the cave, using measurements taken in sondages.

29Since 2003, analyses of the d13C of CO2 present in the water and in the subsurface atmosphere (between ‑21.5‰ and ‑23.5‰) indicate a biogenic origin (Lopez et al. 2007; 2008) corresponding to the disintegration of organic matter by bacterial fauna and root respiration (Emblanch et al. 2003). The prospections realized in the fills by means of dynamic penetrometers (fig. 37) allowed the holes to be equipped with small‑diameter tubes in which CO2 measures can be taken. We have been able to record contents of up to 8.5% in this formation (fig. 38), but the measurements taken in the soil never exceed 2%.

Image 10000000000002FB00000221841CC4757654D5C9.jpg

FIG. 37 – Positioning of penetrometer probes equipped with gas meters for measuring CO2 in the fillings, to the east of Lascaux Cave.

© B. Lopez / université Bordeaux 1, laboratoire Ghymac.

Image 100000000000042C000003ABB925C9DDC0DF5AC5.jpg

FIG. 38 – Diagram of CO2 concentrations measured by penetrometers at different times of year, depending on the spot heights achieved by the surveys: level of the outlet of airlock 1 and variation range of pCO2 in the water (values measured in the soils above the limestone and fills).

© B. Lopez / université Bordeaux 1, laboratoire Ghymac.

30These observations made in the atmosphere of the site of Lascaux and in its waters have led us to propose a diagram of the hydrogeological operation of the epikarst system of Lascaux.

4.3.4 Hydrogeological operation diagram

31The current hypothesis for the hydrogeological operation is that the water present in the cavity originates from drainage of the limestone massif that forms the promontory enclosing the cave, as well as the sand‑clay fill formations. Previous studies have shown that the paleocanyons had no flow limit, due to the lithology of the ground and the depth of these canyons. Our work shows that this slightly permeable fill nonetheless allows a certain quantity of water to flow through. In effect, successive water‑content profiles over time have allowed us to monitor variations that reveal the passage of an infiltration front. In addition, in the very rainy spring of 2008, we indeed observed water in the pipes that traverse the fill, which drained into the limestone massif within a few days. The sections made in the environment of the cave based on topographical, penetrometric and geophysical data indicate that to the east and northeast of the cave, the limestone substratum at the bottom of the fill is situated above the exsurgence zone present in the cave at airlock chamber 1 (fig. 39).

Image 10000000000006570000042989C8358F9C8894CB.jpg

FIG. 39 – Schematic geological sections AB and AC based on interpolation of depth data for the limestone substratum.

© B. Lopez / université Bordeaux 1, laboratoire Ghymac.

32The observations made during our study show that there are some preferential drainage zones in the epikarst of Lascaux, though the flow speeds remain low. The functioning of the system is closer to that of a fissured environment than a karst environment. The system functions only when the humidity level is high enough in the surface and in the epikarst, corresponding to an average refill rate of 250 mm of efficient rain. At this point, when the epikarst groundwater table is created, the pressure is transferred rather quickly—around one to two days after the rain falls. The water that reaches the outlet is pushed by a piston effect. It is never direct rainwater. The epikarst fulfils its role as a sponge, which is very important for the preservation of the cave.

33The fill has a major role in mineralizing the water, since it is home to one—or several—mechanisms responsible for CO2 concentration. It reaches considerable pressure (up to 8%) and then spreads into the limestone, is dissolved in the water or accumulates in the karst trap zones. This explains the very high concentrations reached in the Shaft, as well as the partial CO2 pressure found in the infiltration water.

34The water that reaches airlock chamber 1 presents little direct danger to the art works of Lascaux since it is intercepted and pumped out of the cave. It plays a fundamental role, however, both in the thermal impact as regards the walls, and on the air temperature in the cave. Finally, it provides information concerning the mineralization of the infiltration water that reaches the walls in other manners. In effect, the diagram in figure 40 shows that other flows exist in the cave, which threaten the art works more directly, especially the non‑perennial flows found at the joints of limestone ledges in the cave.

Image 10000000000005690000022A389FE386DA94F179.jpg

FIG. 40 – Diagram showing hydrogeological functioning and CO2 dynamics in the epikarst zone of Lascaux Cave.

© B. Lopez / université Bordeaux 1, laboratoire Ghymac.

4.4 Support properties

4.4.1 Lithological and petrographical description of the formations visible in the cave

35This formation, which belongs to the major series of detritic and biodetritic Coniacian limestones, displays a succession of eight strata of superimposed limestones (Schoeller 1965) that are distinguished by the abundance of detritic intake and their state of diagenetic compaction. Stratigraphically, the succession of sedimentary deposits is disturbed by the presence of inter‑ledges that form open joints, three of which are clearly visible in the first part of the cave. These are generally not very active, but nonetheless constitute the outlet for slow infiltration waters originating from the overlying epikarst aquifer. In the cave, the limestone of the walls is most often compact in the Hall of the Bulls and its axial Gallery. However, due to their initial mineralogical composition and the supergene alteration processes that affect them, these walls can display varying states of superficial cohesion, coarseness and powderiness. This is particularly true for the most sandy, least cemented limestone of the ceiling of the Passageway and some walls in the right gallery, the properties of which are poorly known.

36Few facies have been subject to systematic petrographical and mineralogical studies due to the scarcity of samples that have been taken. With the exception of the late development of white, sub‑automorphic calcite abundantly used by prehistoric humans, analysis of the stone substratum has focused its contrasted morphological properties with the aim of clarifying their implication in the mechanisms of water storage and water transfer. The stone in the cave is a sandy biomicrite (Folk 1962) with sparitic bands and fragments of echinoderms and foraminifers. This compact, stone substratum is heterogeneous (fig. 41a‑b).

Image 10000000000004B0000004BF8B5BFB69561182F0.jpg

FIG. 41 – Start of the Axial Gallery, left wall: a view of polished section of hard, Coniacian limestone. Details: b carbonated cementing of remobilized detritic quartz with sandstone filling; c sedimentation limit between the hard substrate, bottom, and the sandy-clayey iron sections, top (MOP observations; polarized light analysed).

© J.-D. Mertz / LRMH.

37At the microscopic scale, the concentration, distribution and morphology of quartz underlines the importance of the energy implied, on one hand, during the transport of these elements to their sedimentation basin and, on the other, during their subsequent remobilization under the conditions of karst formation (fig. 41ac).

38The stone preserves the signature of multiple syndiagenetic or postdiagenetic episodes (dissolution‑redeposition), such as the filling of non‑continuous animal burrows of varying lengths. Particularly in the Hall of the Bulls, this substratum is covered by carbonated concretions several millimeters thick and of the same nature as the bedrock below. They often appear in the form of scales or crusts that adhere to the substrate. They are affected by the still active dissolution/precipitation processes, which might be explained by the fluctuations in pCO2 described above.

39These “crusts”, which are sometimes coherent and sometimes fragile when they are pierced by filled or open veins (fig. 42), are covered in some places by sandstone and clay or clay and iron deposits. Deposited in thin layers or concentrated on the rough surfaces of the stone, they represent the final events in the development of the cave, although some may be exogenous or anthropic. The corresponding clay phases were analyzed after the fine fraction was extracted on paste oriented by X‑ray diffractometry (Bragg‑Brentano system, Co tube, 1D LinxEye sensor). Interpretation of the the diagrams (fig. 43) shows kaolinite, illite and a majority of calceous smectite that is very sensitive to swelling and shrinking, even with a small variation in the internal climatic conditions (p/po ~ 0.99; T ~ 12°C). This clay succession is fairly constant in the different parts of the cave.

Image 10000000000001D20000015C951E422FC8CC91C2.jpg

FIG. 42 – Partial reprecipitation of calcite in a crevice in scales covering the limestone substrate (MOP observation; non-polarized light).

© J.-D. Mertz / LRMH.

Image 10000000000002FA000003E4499A1C8BC50A1EA7.jpg

FIG. 43 – X-ray diffractogram of clay minerals. Orange: no treatment; blue: ethylene glycol treatment; black: heating; green: hydrazine treatment. Proof of smectite swelling from 12 to 17 angströms.

© J.-D. Mertz / LRMH

40The role played by these clay deposits is determinant in the behavior of the walls since these deposits, which react very strongly due to their large specific surface, have remarkable water‑retention and cationic exchange abilities.

4.4.2 Limestone substratum storage and water transport properties

41The microstructure of the hard, Coniacian limestone has been characterized on a microsample using a destructive technique based on the injection of pressurized mercury (Washburn 1921; Purcell 1949; Pellerin 1980) with an Autopore V Micromeritics porosimeter, HP 210 MPa cell, precision 10‑3 μL).

42The porosity of the stone substratum is approximately 1.5% (fig. 44), with an apparent volume density of 2.62 g.cm‑3 and a low specific surface (0.4 m².g‑1). In comparison, the scales are composed of more open materials, partially detached from the substratum and pierced by a complex network of calcitic veins that have dissolved in some places. This microstructure justifies porosities above 10‑15%, whereas the superficial clay deposits develop significantly higher total porosities (> 30%). There is thus a normal graduation of open porosity from the deep parts of the stone massif toward the surface, which is linked to the cave formation conditions themselves and the modification of the epidermis in contact with the subterranean atmosphere. One remarkable point concerns the dimensions of these pores, or more precisely, the volumes that are open and accessible to fluids for a given pore size. The “healthy” stone is essentially microporous and has a unimodal distribution, concentrated around a pore size of 0.03 μm.

43Based on the considerations of William Thomson (1870) concerning condensation humidity in function of the distribution of the sizes of the pore openings, it seems that, under normal cave humidity conditions (RH ~ 99%), the subsurface of the surrounding stone constitutes a porous environment that is nearly saturated with water solely through endogenous condensation. The application of the predictive model of Katz‑Thompson (1986) to analyze the hydraulic conductivity gives intrinsic permeability values of around 2.10‑15 m², which are characteristic of nearly impermeable systems. Comparing in situ observations (seepage, etc.) with interpretation of the substrate water content measurements confirms that the water saturation in the stone is nonetheless very high and reinforces the fact that the liquid water intake in the cave occurs more through the porosity of connected joints and fissures than through matrix percolation in the strict sense.

44On the cave walls, fluctuations in the internal microclimate associated with the microstructural peculiarities of the different stone facies determine the state of saturation of water that is adsorbed and/or condensed in the subsurface.

4.4.3 Hydric occurrences on the walls and their geochemical properties

45There are very few hydric occurrences, in the form of condensation seepage, visible on the walls of the Lascaux Cave because the thermal and hygrometric regulation equipment is designed to prevent this from occurring. However, between periods when the equipment is used, and due to its inertia, condensation can sometimes appear on the walls and saturate the pores completely; this is what happened in December 2004. One of the objectives of our study was to thus to characterize the phenomenon of condensation using a geochemical tool. Samples were taken using syringes and under the supervision of the cave personnel in order to analyze this water.

46On the walls, hydric occurrences result in drops of water that are particularly concentrated at the joints of limestone ledges. Some of the flowing or dripping water collected on the ledges in the Hall of the Bulls and stagnated in small depressions in the stone. Finally, seepage water, a solution that appears in the form of a gel, has been removed from the sloped surfaces of the Axial Gallery. The major ions in this water were determined by means of liquid chromatography analysis on Dionex 120 equipment. The chemistry of the water collected shows significant abnormalities for many elements (tabl. I).

Image 10000000000004140000010B5AD16A75F0928996.jpg

TABL. I – Main chemistry in mg/l in the condensation and seepage water removed from the bench in the Hall of the Bulls, the sloped surfaces of the Axial Gallery and the walls in the Hall of the Bulls (condensation period: November 2004).

47Before presenting an analysis of the results, we should state that the high concentrations of some elements could partly result from reconcentrations by evaporation since these waters remained in contact with the walls for a certain length of time.

48The main abnormality concerns chloride ions, especially on the ledges in the Hall of the Bulls, where contents of 4 to 5 g required us to realize significant dilutions. In comparison with the flows in the airlock chamber 1 groundwater, which have average concentrations of 5.7 mg/l (tabl. II), we can conclude that the water collected in these three sections is no longer “natural”. We were also very surprised by the nitrate concentrations, which attain considerable values though there is none in the water in the airlock chamber 1 groundwater.

Image 100000000000040B000000B4998D5AC07F5405A0.jpg

TABL. II – Main ion chemistry (in mg/l) in the water originating from leaching by flowing over one of the bench surfaces in the Hall of the Bulls and on the ledge in the machine room on 08/11/2007. Comparison with the mean chemistry of the water collected in a joint in the Hall of the Bulls and in the water from the stratum of airlock 1, section 3.

49The degradation of the quaternary ammoniums used in the treatment of the walls against the spread of the fungi after the work in the cave in 2001 can probably explain these high concentrations. The strong nitrate concentrations come from the oxidation of ammonium when acted upon by nitrifying bacteria.

50We also find significant abnormalities in cations, Ca2+, Mg2+ and particularly K+. As a first approximation, the high Ca2+ and Mg2+ content could be explained by the cationic exchanges between the ammonium in the treatment products and the cations in the sheets of clay that are particularly present on the ledges and sloped surfaces. These mechanisms are quite common and are used for waste storage when passive waterproof barriers must be created. A recent thesis (Gautier 2008) has shown that “the presence of ammonium in interleaf cations reduces the hydrating abilities of the SWy‑2 (NH4) smectite”; a clay that is found at Lascaux. “Studying the interactions using a ‘batch’ method has shown that acetic and formic acids, acetonitrile, dichloromethane, MTBE, benzene, acetone and ethanol are not, or are very poorly, retained by the ammonium clay. On the other hand, chloroacetic and oxalic acids, aniline, formamide and phenol are fixed according to different mechanisms”.

51However, in the cationic stream of smectitic clay at Lascaux that is essentially calciferous, K+ is difficult to exchange, and the use of treatment products cannot account for all the chemical abnormalities identified in the solutions, especially for potassium. The origin of the latter in the condensation water on the walls must therefore be looked for in the dust that is stuck to it (potassic clay?), or in the other pigments (potassic manganese oxide, cryptomelane) used for the paintings (Vignaud et al. 2006), hence the need to avoid these condensation periods using powerful climate assistance equipment.

52Very little water could be collected using syringes, which prevented us from conducting detailed research on the other compounds present in this water, especially NH4 + and some organic compounds that are nonetheless visible on the chromatograms produced by the analysis. If we postulate that the NH4+ ions definitely come from the treatment products, where do these small organic compounds come from? This must come either from the mineralization of organic matter present in the clay or deposited on the walls during countless visits after the discovery, or the degrading of quaternary ammonium that is rich in carbon chains (C12 chains). We understand that these ionic elements (K, NO3‑) and small organic compounds could act as nutrients for living species such as fungi. We have not had the opportunity to encounter the condensation conditions that existed in November 2004. To validate these initial results, we also leached the walls by creating a flow of water on the surface using a distilled water pipette. The water collected thus did not have time to stay in contact with the stone. The results obtained from this leaching, conducted on the ledge of the Hall of the Bulls and on the ledge in the machine room, two unornamented areas that have undergone much chemical treatment, are reported in tabl. II.

53We observe an astonishingly high level of mineralization in this water flow, and more chloride and nitrate abnormalities in comparison to the water from the outlet in airlock chamber 1 but also with the percolations coming from the inter‑ledge joints. These joints flow in very small quantities, but recurrently each year, with a time lag that is sometimes very marked in the groundwater flow through airlock chamber 1. Samples are taken using a porous candle pushed into the joint up to a depth of 50 cm; it sucks the water up to a sampling bottle by means of capillary depression. The high calcium concentrations, despite a very short contact time between the water and the stone, can probably be explained by the dissolution, during leaching, of original or transported calciferous components, the origin of which is still undefined.

54Measures of the dissolved organic carbon taken during leaching indicate values of 4.5 mg/l. These show that nutritional conditions adequate for the development of bacteria or fungi do exist on the walls of Lascaux. The question remains as to which factors triggered the biological explosions observed in the cave since 2001, even before quaternary ammonium was used.

4.5 Conclusions and new questions…

55The studies conducted since 2003 at Lascaux have allowed us to considerably improve our knowledge of the geological and hydrogeological environment of the cave. This research is crucial in the short term, but also in the long term, since it will become an important archive for monitoring the development of the waters in the system and their microclimatological impact on the cave.

56We observe that over the past several years there has been a regular decrease in the hydric refilling of the limestone massif due decreased precipitation and an increase in evapotranspiration. This results in a significant decrease in water flow in airlock chamber 1, which is the outlet of the system. Future work must focus on quantifying this intake and its variation over time. Only modeling the system will allow us to simulate future scenarios in function of global climatic evolution, commonly known as “global warming”. The geophysical and geotechnical studies conducted during our work on the geometry of the different formations will contribute to perfecting a conceptual model of the epikarst system at Lascaux, which will form the basis for all modeling. Another field of investigation concerns CO2. The mechanisms that explain the concentrations measured at the site are not yet well known. It will be necessary to reproduce these processes in a laboratory in order to understand and quantify them. The question of CO2 pumping in Lascaux Cave remains a complex one. It is currently pumped out to ensure the safety of the people who must go into the cave. However, this pumping of carbon dioxide causes a great difference in partial CO2 pressure between the percolated water and the air in the cave. Could this be the cause calcite of crystal growth under the scales on the benches, as well as under the scales on the decorated calcited walls? Finally, some of the chemical measurements of the water collected on the walls of Lascaux Cave are worrying since they prove the anthropogenic nature of the walls. Specific research must be carried out to determine the origin of these nutrients on the walls and their distribution throughout the cave. Is it homogenous or variable? Does the presence of fungi depend on the quantity of nutrients? Specific analyses of the cationic exchanges in the clay, between cations and ammonium, or of small organic chains, will be valuable for characterizing the support at Lascaux and its relationships with the biological activity.

57The studies currently being conducted by the LRMH and the GHYMAC Laboratory of Bordeaux 1 University, on the pore scale and the massif scale, must converge to improve our knowledge and respond to these important questions for the preservation of Lascaux Cave. Acknowledgments The authors would like to thank the Minister of Culture, the DRAC of Aquitaine (Regional Department of Cultural Affairs), the Lascaux Cave Scientific Committee, and everyone who has contributed to our knowledge of Lascaux Cave.

58References

59Bakalowicz 1979 : BAKALOWICZ (M.). — Contribution de la géochimie des eaux à la connaissance de l’aquifère karstique et de la karstification. Thèse doct. d’état sc. nat. : univ. Paris VI, 1979, 269 p. 

60Bourges et al. 2001 : BOURGES (F.), MANGIN (A.) HULST (D. d’). — Le gaz carbonique dans la dynamique de l’atmosphère des cavités karstiques : l’exemple de l’aven d’Orgnac (Ardèche) = Carbon dioxide in karst cavity atmosphere dynamics : the example of the aven d’Orgnac (Ardèche). Comptes-rendus de l’Académie des sciences, série 2A, 333, 11, 2001, p. 685-692.

61Denis et al. 2005 : DENIS (A.), LASTENNET (R.), HUNEAU (F.), MALAURENT (P.). — Identification of functional relationships between atmospheric pressure and CO2 in the cave of Lascaux using the concept of entropy of curves. Geophysical research letter, 32, 5, 2005, p. L05810.1-L05810.4.

62Drogue 1969 : DROGUE (C.). — Contribution à l’étude quantitative des systèmes hydrologiques karstiques, d’après l’exemple de quelques karsts périméditerranéens. Thèse doct. sc. nat., univ. Montpellier, 1969, 462 p. 

63Emblanch et al. 2003 : EMBLANCH (C.), ZUPPI (G.M.), MUDRY (J.), BLAVOUX (B.), BATIOT (C.). — Carbon 13 of TDIC to quantify the role of the unsaturated zone : the example of the Vaucluse karst systems (Southeastern France). Journal of hydrology, 279, 1-4, 2003, p. 262-274.

64Folk 1962 : FOLK (R.L.). — Spectral subdivision of limestone types. In HAM (W. E.) ed. — Classification of carbonate rocks. Tulsa : American association of petroleum geologists, 432, 1, 1962, p. 62-85 (Memoir of AAPG ; 1).

65Gautier 2008 : GAUTIER (M.). — Interactions entre argile ammoniée et molécules organiques dans le contexte du stockage des déchets : cas de molécules à courtes chaînes. Thèse doct., univ. Orléans, 2008.

66Katz, Thompson 1986 : KATZ (A. J.), THOMPSON (A. H.). — Quantitative prediction of permeability in porous rock. Physical review, Série B, 34, 11, 1986, p. 8719-9191.

67Lastennet 1994 : LASTENNET (R.). — Rôle de la zone non saturée dans le fonctionnement des aquifères karstiques. Thèse doct., univ. Avignon et Pays de Vaucluse, 1994, 239 p.

68Lastennet, Mudry 1997 : LASTENNET (R.), MUDRY (J.). — Influence of the flow conditions on the functioning of heterogeneous karstic systems. Environmental geology, 32, 2, 1997, p 114-123.

69Lastennet et al. 1996 : LASTENNET (R.), PUIG (J.M.), EMBLANCH (C.), BLAVOUX (B.). — Influence de la zone non saturée sur le fonctionnement des systèmes karstiques : mise en évidence dans les sources du Nord-Vaucluse. Hydrogéologie, 4, 1996, p. 57-66.

70Lastennet et al. 1999 : LASTENNET (R.), DENIS (A.), MALAURENT (P.), VOUVE (J.). — Behaviour of the epikarstic aquifer: signal analysis and flow analysis. Site of Lascaux Cave. In ANDREO (B.), CARRASCO (F.), DURAN (J.J.) ed. — Contribucion del studio cientifico de las cavidades karsticas al conociiento geologico. Malaga : Patronato de la cueva de Nerja, 1999, p. 363- 370.

71Lopez et al. 2007 : LOPEZ (B.), LASTENNET (R.), MALAURENT (P.), DENIS (A.). — Characterization of seepage waters in a filled epikarst. In INTERNATIONAL ASSOC. OF HYDROGEOLOGISTS ed. — Groundwater and ecosystems. IAH congress (XXXV ; Lisbon ; 2007), p. 353-355. Résumés des communications et articles sur CD.

72Lopez et al. 2008 : LOPEZ (B.), LASTENNET (R.), MALAURENT (P.), DENIS (A.). — The recharge features in karstic context : a case study from the Lascaux cave, France. In INTERNATIONAL ASSOCIATION OF HYDROGEOLOGISTS ed. — Integrating groundwater science and human wellbeing. IAHS congress (XXXVI ; Toyama ; 2008).

73Mangin 1975 : MANGIN (A.). — Contribution à l’étude hydrodynamique des aquifères karstiques. Thèse doct. d’État, univ. Dijon, 1975, 124 p. 

74Mudry 1987 : MUDRY (J.). — Apport du traçage physico- chimique naturel à la connaissance hydrocinématique des aquifères carbonatés. Thèse doct. sc. nat., univ. Besançon, 1987, 381 p. 

75Pellerin 1980 : PELLERIN (F.). — La porosimétrie au mercure appliquée à l’étude géotechnique des sols et des roches. Bull. de liaison du labo. des Ponts et Chaussées, 106, 1980, p. 105-116.

76Peyraube et al. 2008 : PEYRAUBE (N.), LASTENNET (R.), LOPEZ (B.), SIRIEIX (C.), NAUDET (V.), DENIS (A.). — Epikarstic hydrogeological reconnaissance using geophysical methods : case study of the Lascaux cave (France). In INTERNATIONAL ASSOCIATION OF HYDROGEOLOGISTS ed. — Integrating groundwater science and human wellbeing. IAHS congress (XXXVI ; Toyama ; 2008).

77Purcell 1949 : PURCELL (W.R.). — Capillary pressures : their measurement using mercury and the calculation of permeability therefrom. Journal of petroleum technology, 1949, 186, p. 39-47.

78Shoeller 1965 : SCHOELLER (H.). — Étude géologique, hydrogéologique et climatologique de la grotte de Lascaux pendant le cycle 1964-1965. Rapport pour la première commission scientifique internationale de la grotte de Lascaux, univ. Bordeaux 1, 1965, 97 p. 

79Spötl et al. 2005 : SPÖTL (C.), FAIRCHILD (I. J.), TOOTH (A. F.). — Cave air control on dripwater geochemistry : Obir Caves (Austria) : implications for speleothem deposition in dynamically ventilated caves. Geochimica et cosmochimica acta, 69, 10, 2005, p. 2451-2468.

80Thomson 1870 : THOMSON (Lord Kelvin W.). — On the equilibrium of vapour at a curved surface of liquid. Proceedings of the Royal society, 2, 1870, p. 63-70.

81Vignaud et al. 2006 : VIGNAUD (C.), SALOMON (H.), CHALMIN (E.), GENESTE (J.‑M.), MENU (M.). — Le groupe des « bisons adossés » de Lascaux : étude de la technique de l’artiste par analyse des pigments. L’anthropologie, 110, 4, 2006, p. 482-499.

82Vouvé 1968 : VOUVÉ (J.). — Colline de Lascaux (commune de Montignac) : détermination d’un modelé karstique original en pays calcaire. Comptes-rendus de l’Académie des sciences de Paris, 266, 1968, p. 2139-2142.

83Washburn 1921 : WASHBURN (E.W.). — The dynamics of capillary flow. Physical review, 17, 1921, p. 273‑283.

84Williams 1982 : WILLIAMS (P.W.). — The role of the subcutaneous zone in karst hydrology. Journal of hydrology, 61, 1982, p. 45-47.

Notes de bas de page

1 Funding granted to Benjamin Lopez, 2005-2008.

Précédent Suivant

Le texte seul est utilisable sous licence Licence OpenEdition Books. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.