1Over the last 20 years, great progress has been made in the field of karstogenesis and in the development of multidisciplinary approaches to deciphering the palaeoenvironmental information contained within karst forms and formations (Delannoy et al., 2009). Alpine karsts have played an important role in these studies, most notably through the impetus of an international working group that was set up in 2000 at the 'Cave Genesis in the Alpine Belt' workshop in Habkern, Switzerland (Häuselmann and Monbaron, 2001). Comparisons of the methods used and the results obtained from one end of the Alpine chain to the other have provided an overview of the state of knowledge of Alpine cave genesis. It also enabled workers to identify and fill gaps in this knowledge, and suggested avenues for new or further research, while retaining as a guiding principle and common denominator the decryption of the information contained in the caves of the Alps (Audra, 2004; Audra et al., 2006, 2007). This information can be categorised into three main types of indicators and records:
2(i) Geometrical (Goran, 1992): The position and three-dimensional structure of cave systems can be determined by applying appropriate analytical techniques (Hobléa, 1999b) to data on the most voluminous parts of these cave systems, which are accessible to, and have been mapped by cavers. The results of such studies may then be combined with indicators such as palaeoflow paths (i.e., 'chemin de drainage', Choppy, 1994. We consider here the palaeoflow path as dominant direction of karst drainage determined by the location of the input and the emergence; this direction may change from one phase of karstification to another) and different tiers of passages (Audra et al., 1993), in order to identify the main stages in polyphase cave genesis (Palmer, 1991);
3(ii) Morphological: Feature of the walls, ceilings and floors of passages provides information about their genesis (phreatic, vadose or epiphreatic regime, flow direction, phases of aggradation or incision, gravity movements, etc.);
4(iii) Sedimentological: Sediment fillings in mountain karsts are extremely important palaeogeographic archives and indicators (Audra, 1995). Numerous techniques have been developed for analysing detrital and chemical (speleothems) deposits, thereby enabling them to be used as records of variations in continental environments, alongside other natural archives (Sasowsky and Mylroie, 2004). Furthermore, cave deposits are the only parts of caves that can be dated directly, as the passages themselves can only be dated with reference to the age of the host rock (upper limit) and the age of the oldest sediment deposits in cavities (Häuselmann, 2007). Dating is needed in order to produce a chronology for the identified karstification phases. Of the numerous methods that have been developed, the most commonly used is U/Th isotope dating of the calcite in speleothems. However, a major drawback of this method is that it cannot be used to date samples older than 350 ka (end of the Middle Pleistocene), and U/Th analyses of speleothems in numerous studies of Alpine karsts have indicated ages far beyond that range (Maire, 1990; Audra, 1994; Delannoy, 1998; Hobléa, 1999a; Häuselmann, 2002, etc.). To overcome this limitation, karst scientists in the Alps have applied a number of other approaches. At first these approaches focused on relative dating methods such as palaeomagnetism (Audra and Rochette, 1993; Hobléa, 1999a) and palynology (Hobléa, 1999a); however, the development of cosmogenic nuclide methods applied to geomorphology in 1990s (Siame et al., 2000, 2006) opened new and interesting possibilities for dating cave sediments (Stock et al., 2005), especially those shown to be older than the limits of the U/Th dating method on speleothems. Initially developed to date periods of surface denudation and incision, the cosmogenic nuclide method was later adapted for dating mountain cave deposits, first by Granger (Granger et al., 1997, 2001) in the USA, and then by P. Häuselmann, aided by D.E. Granger (Haüselmann and Granger, 2004, 2005), in Europe. This method was tested for the first time in France in 2005, at Mount Granier, in the sub-alpine Chartreuse Mountains. The tests were carried out as part of a larger, international research project on alpine cave genesis. The present article describes the geological setting of these cave sediments, together with the methods used, the difficulties encountered, the results and their palaeogeographical implications.
5Mount Granier lies at the northern extremity of the perched syncline that forms the eastern edge of the sub-alpine Chartreuse Mountains (fig. 1) and which contains the Hauts de Chartreuse Nature Reserve. This perched syncline runs north-south from Mt. Granier to the Dent de Crolles, and is divided into several sub-units by transverse strike-slip faults.
Fig. 1 – Location of Mount Granier at the northern extremity of the Hauts de Chartreuse Nature Reserve (Chartreuse Massif, France).
Fig. 1 – Localisation du mont Granier à l’extrémité nord de la Réserve Naturelle des Hauts de Chartreuse (Parc Naturel régional de Chartreuse, France).
6Unlike the sub-units to the south, which form hanging valleys, the Mt. Granier is a perched half-syncline that runs from an altitude of 1933 m on the western crest, to 1578 m at the summit of the eastern rim, from where the truncated synclinal trough can be seen (fig. 2). Having lost its eastern flank, the synclinal plateau slopes east towards the Isère Valley (Grésivaudan), following the 10° to 15° dip of its constituent Neocomian series, which is crowned by a 200-m thick sequence of Urgonian limestone (a large part of the Upper Urgonian has been removed by erosion). The north face of the Mt. Granier, which was formed by a huge landslide in 1248, overlooks the Chambéry-Montmélian valley, which divides the Chartreuse from the Bauges Mountains.
7The summit of Mt. Granier consists of a narrow 2.8-km2 plateau that is elongated along its central axis. This plateau clearly stands out from the surrounding landscape (fig. 3) and dominates the surrounding valleys and neighbouring plateau, from which it is separated by a south-facing fault escarpment. The highly-fractured Urgonian limestone of Mt. Granier, that represents a volume of 0.75 km3, has been strongly affected by gravity decompression due to the current perched position of the massif. In addition, Mt. Granier is subject to a cool, humid climate (mean annual temperature at 1700 m: 2.5°C) and humid (mean annual precipitation at 1700 m: 2500 mm). Soils and vegetation are mostly typical of the mountain and sub-alpine stages, and include both forest (softwood) and herbaceous cover.
Fig. 2 – The perched half-syncline of Mount Granier.
Fig. 2 – Le volet synclinal perché du mont Granier.
1: Berriasian marls; 2 & 3: Valanginian limestone; 4: Hauterivian marls; 5: Lower Barremian marly limestone; 6: Barremian limestone (Urgonian facies).
1 : marnes berriasiennes ; 2 & 3 : calcaires valanginiens ; 4 : marnes hauteriviennes ; 5 : calcaires marneux du Barrémien inférieur ; 6 : calcaires à faciès urgonien du Barrémien.
Fig. 3 – Aerial view of the perched half-syncline of Mount Granier.
Fig. 3 – Vue aérienne du volet synclinal perché du mont Granier.
Photo: C. Kerkhove.
Cliché : C. Kerkhove.
8The characteristics of the Mt. Granier and the neighbouring mountains of the Hauts de Chartreuse are highly favourable to karstification - a gently inclined, tabular volume of pure limestone receiving abundant quantities of cold precipitation that percolates through soils that have been acidified by vegetation growing on a highly permeable substrate. Present denudation rates for the Mt. Granier, calculated using Corbel’s formula, as adapted by B. Hakim (1984), are about 90 mm/ka, or 250 t.km-2.a-1 (Hobléa, 1999a). Hence, it is not surprising that, like the Vercors to the south, the Chartreuse contains extremely well-developed cave systems of underground passages. However, this development has taken on exceptional proportions in the Chartreuse, as the Hauts de Chartreuse area has the highest density of mega-cave systems than anywhere else in the Alps, and one of the highest densities of cave passage per volume unit in the world. Within the Hauts de Chartreuse Nature Reserve, the Mt. Granier (Hobléa, 1999a) just slightly exceed the Dent de Crolles (Lismonde et al., 1997) for the record of highest cave density, having more than 55 km of mapped, interconnected passages under a surface of less than 3 km2. The Mt. Granier also includes secondary cave systems that are not connected to the main system by cavers exploration, bringing the total extent of the cave systems within the mountain to 90 km (fig. 4). According to R. Maire (1980), the ratio of the cave volume to the karstified rock is about 1.3.10-3. This figure, similar to that for the Dent de Crolles (1.1. 10-3), is orders of magnitude higher than the usual values for surrounding areas (e.g., n.10-5 in the Bauges). On the surface, the density of the caves is reflected in the high density of known entrances, exceeding 110 caves/km2 (Durand and Nant, 1998).
Fig. 4 – Map of Mt. Granier’s cave system.
Fig. 4 – Carte spéléologique du massif du Granier.
Spéléo Club de Savoie, 2005, unpublished.
Spéléo Club de Savoie, 2005, inédit.
9The vastness of the cave system could be explained by present karstification conditions: at the current rate of chemical erosion (90 mm/ka, see above) and an ablation rate of 20% of total specific dissolution, almost 20 ka would be needed to form the Mt. Granier cave system (approximately 1 Mm3 of known, penetrable karst voids). This already takes us beyond the Holocene. However, even if the remaining 80% of the specific dissolution is dedicated to sub-surface ablation, more than 1.2 Ma would be needed to erode the 90 m of Urgonian limestone missing from the Mt. Granier (excluding glacial overdeepening in the cirque in the middle of the plateau), although this remains within the Quaternary. Thus, in theory, the karstogenesis of the Mt. Granier could have occurred entirely within the Quaternary, as was long thought to be the case for most of the cave systems in the Alps and pre-Alps. The tendency before the 1990s was indeed to minimise the importance of pre-Quaternary karst heritages (reduced to pre-Tortonian palaeokarsts) in favor of the action of glacially derived flows (Chardon, 1981, 1984, 1989).
10Following a reevaluation of the pioneering work of C. Mugnier (1965), one of the first scientists to take an interest in the Mt. Granier, combined with increased knowledge of the role of glaciers and glaciations in karstification (Ford, 1987, 1992), we began, along with other karst researchers in the 1990s (Bini et al., 1997), to reconsider the then dominant idea that cave systems were primarily the product of glacial meltwater. Also at this time, exploration of the Mt. Granier’s cave system reached a turning point with the 1988 discovery of the Balme à Collomb Cave and its internationally important deposits of cave bear fossils. The length of passages explored increased exponentially between 1988 and the early 2000s, and the area became an important field site for understanding the genesis of underground mega-cave systems and their relationship with changes in regional palaeogeography.
11'Speleographic analysis' of the Mt. Granier underground system (Hobléa, 1999b) revealed the existence of five tiers of passages in this extremely complex, three-dimensional labyrinth (fig. 5). These tiers were formed at or near former fluvial base levels, and are now perched (Audra et al., 1993). Each tier occupies a specific altitude range (fig. 6). All of the tiers slope eastward, generally quite steeply, with inclination that increase from one tier to the next, from highest to lowest. Each tier has a substantially different flow path (Choppy, 1994). An analysis of the whole system reveals a gradual clockwise rotation (NE toward E) of successive flow paths (fig. 7). The tiers are also different, regarding both morphology (fig. 7) and sediments (Hobléa, 1999a). They are perforated and interconnected by shafts and vadose invasion shafts, indicating periods of incision/verticalisation and the capture of flows from a higher tier toward a lower tier.
Fig. 5 – 3D view of the five cave tiers in Mt. Granier, showing steeper gradient for the lowest tiers.
Fig. 5 – Vue 3D des 5 étages de galeries karstiques du mont Granier, montrant un gradient vertical croissant pour les étages inférieurs.
After Spéléo Club de Savoie.
D’après Spéléo.Club de Savoie.
Fig. 6 – Altitude ranges of the cave tiers in Mt. Granier.
Fig. 6 – Tranches altitudinales occupées par les étages de conduits karstiques du mont Granier.
Fig. 7 – Flow directions and morphotypes of the Mt. Granier cave tiers.
Fig. 7 – Directions de drainage et morphotypes des étages de galeries karstiques du mont Granier.
12Studies combining the geometric structure of the cave system, passage morphology and cave sediment archives (grain size, mineralogy, pattern of deposits, calcimetry, colorimetry, U/Th isotope dates, palaeomagnetism, palynology, palaeontology) have identified three main phases of karstification related to different palaeogeographic and palaeoenvironmental settings. In descending chronological order from uppermost to lowest tiers, it is possible to distinguish:
13(i) A cave-tunnel period, which corresponds to the two highest tiers (E1 and E2). The passages in these tiers drained NNE towards a palaeospring in the Chambéry-Montmélian valley. In the Miocene this valley contained the 'Chambéry River' (Debelmas, 1995), which flowed from the external crystalline massifs, which were then being uplifted. This configuration indicates that excavation of the Isère Valley had not yet removed the eastern side of the Mt. Granier syncline. Tiers E1 and E2 mostly consist of gently-inclined phreatic passages near the top of the limestone mass. E1 lies at around 1800 m, whereas the large palaeomaster caves of E2, including the Balme à Collomb cave bear gallery, lie between 1720 m and 1600 m. Sedimentary fill deposits contain minerals and materials derived from surface regoliths that indicate a warm and humid climate (kaolinite, montmorillonite, remains of iron crusts, etc.). These materials were transported into the passages at the same time as pebbles which indicate the presence of an Upper Cretaceous or Palaeogene calcareous-sandstone cover. This cover has now entirely disappeared from the surface of the Mt. Granier, but Upper Cretaceous outcrops can still be found in the neighbouring hanging valleys of the Hauts de Chartreuse. Further support for C. Mugnier’s (1965) conclusions and the hypothesis of a pre-Quaternary age for these passages is provided by the following observations: dates of the concretions sealing these fills are nearly all beyond the limits of the U/Th method, and some isotope ratios for tier E2 suggest ages of at least a million years; some samples from detrital and chemical deposits (ferrous oxides trapped in concretions) show palaeomagnetic reversals.
14C. Mugnier, who only had access to fragments of the cave network, hypothesised that the passages formed before the last major phase of folding and uplift in the western (external) part of the Alpine chain. In fact, the geometry and morphology of the passages and fillings do not provide any evidence of syn- or post-genetic tilting, and it is now known that the oldest phases of passage formation post-date this phase (Hobléa, 1999a, 2001), which is attributed to the Upper Miocene (around 5.3 Ma). Flow path orientations clearly show a west-east component. Whereas C. Mugnier (1965) envisaged a palaeoexsurgence in the entrance gallery of the Balme à Collomb, on the west face of the Mt. Granier, which was the only part of the network known at the time, the entrance zone for this cavity is now interpreted as an upstream part of a palaeoconduit that drained eastward. Hence, these features are probably palaeodrains truncated by the erosion of the surrounding rock. The cavities seen today are merely the remnants of a system that riddled a larger 'Mount Granier', which extended much further west, north and east (presence of the eastern flank of the syncline) than the current remnant. This system functioned as a binary karst fed laterally by rivers penetrating the plateau.
15By deduction after crossing geometrical, morphological, sedimentological and different types of chronological data, the genesis of these upper tiers of the Mt. Granier cave system could therefore be situated between 5.1 Ma (known upper limit) and the end of the Pliocene (Hobléa, 1999a; 2001).
16(ii) A large underground canyon period. Tier E3 is perched within the limestone mass of the Mt. Granier between 1630 m, for the upstream heads of the network, and a palaeowater table with distinct phreatic tubes that meander between 1490 m and 1450 m. The upstream ends of the network are connected to the overlying E2 tier, from which they capture flows through very steep and imposing canyons. These canyons contain reworked fill deposits derived from tier E2 (pebbles and gravels with sandy matrices). This morphosedimentary association indicates a dynamic setting of rapid lowering of the regional base level, possibly associated with the uplift of the massif and/or the incision of the valley that served as the base level for the karst system. In the 1450-1550 m altitude range, these deposits are overlain by laminated silts, sealed by a concretionary deposit that is older than the limits of U/Th dating. These silts, which are typical of decantation deposits formed in a submerged paraglacial setting, indicate that the passages of this tier had already formed before the Middle Pleistocene and may be much older. The E3 flow path had a distinct eastern component (fig. 7), although the corresponding spring lay further north than those for tiers E4 and E5, in a geomorphological setting marked by the partial ablation of the eastern flank of the Mt. Granier syncline.
17(iii) A meanders-shafts period, from Middle Quaternary to Recent, correlates with tiers E4 and E5. These tiers are mostly composed of shafts and meandering galleries and passages (narrow, sinuous underground canyons that are higher than they are wide) that mostly formed under vadose conditions, although they were periodically subject to phreatic conditions due to blocking of the exsurgence during glacial maximums. The subsequent flooding of the passages in these tiers deposited fine sediments (laminated carbonate silts). These two superposed tiers formerly drained (tier E4) and currently drains (tier E5) toward emergences at the foot of the east side of the Mt. Granier (Isère Valley side), directly below the eastern promontory of the plateau. This seems to show that the eastern flank of the Mt. Granier syncline was already absent and, consequently, that the Isère Valley was already distinct and deeply incised. The massif’s current emergence, the Eparres spring, lies within the Valanginian limestone, at an altitude of 950 m. Water from the Mt. Granier’s underground master cave, which runs through part of tier E5, crosses the Hauterivian marl aquiclude via a fault with a very high hydraulic gradient, at around 1190 m, near the foot of the east rim of the Mt. Granier. According to morphologic clues (shape and slope of the galleries), tier E4 shows a distinct palaeowater table at an altitude of about 1350 m. This level, which is marked by a joint-controlled maze of small tubes, had already been excavated at isotope stage 6 (U/Th dates around 154 ka, Lower Riss). U/Th dates on speleothems of tier E5 show that the Urgonian parts had already been excavated by 33 ka (Hobléa, 1999a, 2001).
18Before the development of the cosmogenic nuclide method, it was impossible to determine an absolute date for the genesis of the upper tiers whom speleothems give ages over the U/Th method limits (E1 to E3). The cosmogenic nuclide dating technique was adapted for endokarst deposits by D.E. Granger et al. (1997, 2001), and then applied to Alpine cave deposits by P. Häuselmann and D.E. Granger (2004, 2005) in the upper levels of the Siebenhengste cave system (Switzerland). Thanks to the presence of quartz sand layers in the fill of the E2 tier, we tested this method in 2004-2005 as part of the comparative studies being carried out by an Alpine cave genesis research group. The experimental protocol and the materials dated are described below (Hobléa and Häuselmann, 2005).
19Two of the six in situ cosmonuclides commonly found are suitable for dating geomorphological phenomena: Beryllium 10 (10Be) and Aluminum 26 (26Al), which are produced by spallation reactions in the nuclei of silica atoms (amongst others) in rock bodies at or just below the surface and particularly in quartz grains.
20It is such quartz grains, contained in the sandy fractions of certain cave deposits that we used to date the formation of the surrounding karst passages. As long as the quartz is exposed to cosmic radiation (i.e., situated on or just below the Earth’s surface), it will produce the radioactive cosmonuclides 10Be and 26Al (half-lives of 1.34.106 a and 7.2.105 a, respectively). When quartz grains are transported underground into caves, they are protected from cosmic radiation, thereby blocking the production of new cosmonuclides. The initial (inherited) cosmonuclide content of the two isotopes will then gradually decay at different rates. Consequently, burial dates for quartz grains can be calculated by measuring their 26Al /10Be ratios. Although the production rates of these two isotopes vary greatly with latitude, altitude, and the presence or absence of soil or snow cover, the ratio between the Al and Be contents of minerals on the surface is constant at about 7:1. However, because the half-life of 26Al is much shorter than that of 10Be, this ratio starts decreasing as soon as the sediment is washed underground. Therefore, as long as a sediment was initially on the surface for a long time enough to become 'charged' with cosmogenic nuclides, a good indication of the date when it was buried in the karst can be obtained. This method can be used to calculate burial dates over a range of approximately 100 ka to 5 Ma.
21Although the principle is relatively simple, analysing the samples and measuring the isotope contents, using accelerator mass spectrometry (AMS), is a long, complex and delicate process that is subject to numerous difficulties and obstacles (Dupuis et al., 2003). The following section describes the protocol used to analyse the Mt. Granier samples (American method; Häuselmann et al., 2003).
22Samples must be taken at least 20 m inside the cave from the entrance, and at least 10 m below the surface, in order to eliminate the possibility of in situ production of cosmonuclides and to limit the risks of contamination from the surface. The sediment must contain a high proportion of quartz grains larger than 0.2 mm (minimum size for standard analyses).
23It requires several steps, including:
- Extraction of the quartz. Organic matter and other minerals are eliminated by a complex cleaning process involving washing, screening, decalcification with HCl or HNO3, dissolution of certain metallic elements using aqua regia, and rinsing with distilled water. The 0.5-1 mm fraction (possibly including the 0.25 mm to 0.5 mm fraction if quantities are otherwise insufficient), is then separated and magnetic elements are removed. Muscovite is eliminated using oxygenated water, and the remaining separate (40 g) undergoes ten 24-h baths in a mixed solution of HF and HNO3, with treatment in an ultrasound bath.
24- Dissolution and concentration of the separated and purified quartz fraction. The quartz is dissolved in Teflon test tubes in a calibrated bath of HF and HNO3 with progressive heating over several days to 70°C. Complete dissolution takes about two weeks, with a finishing process involving H2SO4 and HCl, and centrifuging. After elimination of Fe by chromatography and of Ti by selective precipitation, all that remains is a concentrate of Al and Be.
25- Separation of the Al and Be in this concentrate using chromatography and precipitation. The separate elements are then oxidised and mixed with the metal for spectrometric analysis.
26The presence of in situ cosmogenic nuclides shows the surface origin of the material in the cave deposits. If the spectrometry measurement gives a null result for cosmogenic nuclides, either the sedimentary material was never on the surface, or it has been buried for longer than the method’s measurement range (around 5 Ma). Morphological analysis of the quartz grains can be used to differentiate between these two scenarios. The first element to be measured is 10Be, which has the longer lifespan. 26Al is measured next, as it is more likely to occur in smaller quantities in a sample that has been buried for a long time (starting hypothesis). In order to check the validity of the 10Be measurement (margin of error of the method reported as 1 ), comparative measures of 10Be/9Be ratios are performed on laboratory blanks and the sample to be dated (9Be is added artificially). For our cosmogenic nuclide dating tests, we took eight samples from four different cavities, but only two contained enough quartz (but not enough for running duplicates). These samples came from two widely separated cavities (fig. 4), the Trou Lilou, which is in the central part of the plateau, and the Cuvée des Ours, which drains the northern part of the massif. Most of the passages explored in both caves are palaeodrains and palaeomaster caves of tier E2.
27The Trou Lilou sample was taken from a stratigraphic reconnaissance core drilled during the speleokarstological study of the massif (Hobléa, 1999a). The core was drilled at an altitude of 1705 m in the main gallery - a vast, fossil phreatic conduit that cavers refer to as the 'Megalerie' (fig. 8). The base of the core consists of several meters of detrital material (bedrock was not reached by the borehole). This gallery is now only 30 m below the topographic surface of the plateau, but was covered by bigger one thickness of rocks when it developed along a water table according to a former base level. The core section (fig. 9) shows a stratified succession of decimetre-scaled, non-carbonate sandy layers with a large percentage of fines (silts and clays). The uppermost sandy layers also contain gravel, pebbles and a few angular blocks of limestone that fell from the ceiling. The sequence is covered by a layer of wet moonmilk (pasty calcite). In places, a few isolated stalagmites seal the sedimentary sequence. These stalagmites have not been subject to U/Th dating, because of the expected great ages of the tier and of its sediments. The sample selected for cosmogenic radionuclide dating was taken from layer I LIL 3 (fig. 9). It is an interbedded in layer I Lil 2 (non-carbonate brown sand), which consists of inclined layers of yellowy-green silt (47%) and sand (46%), with black grains (iron oxides). These non-carbonate layers are rich in montmorillonite (marker of warm climates) and illite, and in shiny, sub-angular quartz grains. This morphoscopic characteristic of the quartz grains is found in all the layers of this stratigraphic section and indicates fluvial deposition in the gallery.
Fig. 8 – The 'Mégalerie' in the Trou Lilou: part of the E2 tier, its filling of which was dated to 4.3 Ma using cosmogenic nuclides.
Fig. 8 – La Mégalerie dans le Trou Lilou : élément de l’étage de galeries E2 du Granier, dont le remplissage a été daté par la méthode des nucléides cosmogéniques.
Photo: E. Sibert.
Cliché : E. Sibert.
Fig. 9 – Location and stratigraphic setting of sample ILIL3 (Trou Lilou, Mt. Granier).
Fig. 9 – Situation et contexte stratigraphique de l’échantillon ILIL3 du Trou Lilou (mont Granier).
A: Speleothem. 1: bed of wet moonmilk; 2: non-carbonate sand; 3: interbedded inclined beds of sandy silts (sample ILIL3); 4: gravelly sand; 5: silty sand, non-carbonate; 6: black sand, non-carbonate; 7: non-carbonate sand with thin beds of clay.
A : Stalagmite. 1 : mondmilch hydraté ; 2 : sables non carbonatés ; 3 : lits inclinés en inclusion de limons sableux non carbonatés (échantillon daté ILIL3) ; 4 : sables graveleux ; 5 : sables limoneux non carbonatés ; 6 : sables noirâtres non carbonatés ; 7 : sables à passées argileuses non carbonatés.
28The sample from the Cuvée des Ours was taken at an altitude of 1648 m, in the 'Galerie des Pulsatiles', which is a tributary of the palaeomaster cave. The Cuvée des Ours is a composite gallery that consists of a former phreatic tube that was overdeepened to form a canyon. The fill deposits occur on benches corresponding to remnants of the original floor of the tube (fig. 10). The canyon entrenchment has naturally exposed a section through the fill, which consists of two sequences of detrital deposits, separated by calcite deposition. These two sequences are composed of a sandy layer lying on a fine-grained layer. The fine-grained layer in the upper sequence has a high carbonate content (similar to the hydrated moonmilk layer in the I LIL layer of the Trou Lilou, see above). A thin layer of decarbonated sandy silt is interbedded between the sand of the lower sequence and the calcite layer separating the two sequences. The age of these speleothems exceeds the limits of the U /Th method. The sample for cosmonuclide dating was taken from the sands of the PULS 4 layer (fig. 9). PULS 4 is a gravelly sand with a low carbonate content (11.5%), for which preliminary analyses showed approximately 10% quartz in the < 2 mm fraction. The in situ sediment contained inclusions of fine, grey-brown conglomerate, mostly consisting of iron oxides, and which resembled inclusion I LIL 3, from which the Trou Lilou dating sample was taken. Nearly all the clay in PULS 4 was montmorillonite, which indicates a warm climate, and minor quantities of chlorite. The upper layers were richer in halloysite (PULS 2) and kaolinite (PULS 1), which also indicate primary surface alteration in a warm and humid tropical climate.
29Thus, the two usable samples remaining after the initial processing came from well-separated deposits that show certain stratigraphic and sedimentary similarities. These similarities increase the validity of the resulting dates as possible evidence for a correlation between distant sequences within a single tier of passages.
Fig. 10 – Location and stratigraphic setting of sample PULS4 (Cuvée des Ours, Mt. Granier).
Fig. 10 – Situation et contexte stratigraphique de l’échantillon PULS4 de la Cuvée des Ours (mont Granier).
1: silty sand; 2: white carbonate-rich clay (moonmilk-like); A: speleothem; 3: sandy silt; 4: gravelly sand; 5: grey wet silty clay.
1 : sables limoneux ; 2 : argiles blanches carbonatées (type mondmilch). A : plancher stalagmitique ; 3 : limons sableux ; 4 : sables graveleux (échantillon daté PULS4) ; 5 : argiles limoneuses gris brunâtre très humides.
30Although the isotope concentrations were low (tab. 1), they were sufficient for geochemically reliable measurements, and therefore to obtain ages with a relatively small error range, despite the fact it was not possible to run duplicates.
31The resulting ages of 4.3 Ma for the Trou Lilou and 3.4 Ma for the Cuvée des Ours showed that the sampled sediments in the upper speleological tiers of the Mt. Granier were deposited during the Pliocene. This result validates the age deduced by karstogenesis studies of the massif (Hobléa, 1999a, 2001). We can therefore state that the first drainage structures that are still identifiable in the Mt. Granier formed between the Upper Miocene (after the last major uplift phase in the Alps) and the Upper Pliocene, as the passages of tier E2 (and a fortiori those of E1) were excavated 4.3 Ma to 3.4 Ma ago. The burial age of sample ILIL3 is similar to the age obtained for the uppermost networks of the Siebenhengste in Switzerland (sample SHP7, Häuselmann 2007).
32Given that the 1- errors on these ages are 270 ka, and that 2- errors would exceed 500 ka, the difference of about 850 ka in the two dates obtained for the Mt. Granier samples shows that it is difficult but possible to suggest a chronological correlation between sedimentary sequences sampled in two separate cavities, despite the similarities between the layers (see above). For example, in the Siebenhengste, the ages obtained for samples from two superposed layers in the same sedimentary section differed by 2 Ma (sample SHP7: 4.4 Ma; SHP2: 2.36 Ma, ibid), although an erosion phase was identified between them. The difference in ages for the Mt. Granier samples can also provide information about the times at which of these today fossil upper levels were active, at least during a large part of the Pliocene.
33Combining these cosmogenic nuclide dates with other palaeogeographic and palaeoenvironmental data from speleographic and morphosedimentological analyses suggests that the Mt. Granier was not yet a mountain in the Upper Pliocene; rather, it was a much more extensive plateau bounded by open valleys that mostly extended eastwards towards the present Isère valley. The Cuvée des Ours and the Trou Lilou would have been left bank tributaries on the of a master cave located at the base of the syncline, whose outlet was in the Montmélian-Chambéry valley (fig. 11). This master cave was completely eroded away during the Quaternary, along with the entire eastern flank of the mountain.
Fig. 11 – Schematic reconstruction of the morphological setting of the Mt. Granier at the end of the Neogene, compared with the current situation.
Fig. 11 – Reconstitution schématique du contexte morphologique du paléomont Granier à la fin du Néogène comparée à la situation actuelle.
34To estimate the 'inherited denudation rate' (tab. 1) for the catchment, the initial 10Be isotope concentration (obtained by reintegration of time below ground) was divided by the annual cosmogenic nuclide production rate at the surface (assumed to be 17.3 at.g-1.a-1). It gives an idea of the surface weathering activity during the time before the sediments were deposited underground. This rate is different from the incision rate that measures the rhythm of the entrenchment of the surrounding valleys and that it not yet calculable with precise values for Mt. Granier area. This gives the number of years the sediment remained at or just below the surface, from which the denudation rate can be estimated. However, the result must be viewed with caution, as the values obtained can only be approximations because the absolute (sub)surface altitude of the sediment is not known and the accumulation rate on the Earth varies.
35We obtained values of 7 m/Ma (sample ILIL3; tab. 1), which are very low for the Alps and more typical of rates for stable intracontinental cratons. Could the results obtained for the Mt. Granier indicate the last moments of a Miocene peneplain before the orogenic crisis? Whatever the case, the erosion rate then increased rapidly, increasing by a factor of two within 0.85 Ma (sample PULS4), and more then, with the change to a colder and more mountainous climatic and morphological setting [comparison between the last Pliocene denudation rate (14 m/Ma) and the actual one (90 m/Ma) gives a factor of about six]. During the Pleistocene, this more intense erosion reached and attacked the Urgonian limestone, which had by then been stripped of its Upper Cretaceous (and possibly Tertiary) cover between 1 Ma and 1.5 Ma ago (Hobléa, 1999a).
Tab. 1 – Results of the 26Al and 10Be analyses of the two Mt. Granier samples.
Tab. 1 – Résultats des mesures de concentration en 26Al and 10Be dans les 2 échantillons du mont Granier.
N.B.: The values were standardised for an average life of 1.34 Ma, with 1 sigma as the uncertainty values (Measurement of 10Be contents: Dr. P. Kubik, PSI/ETH, Zürich; measurement of 26Al contents: Lawrence Livermore National Laboratory California, USA; geochronological calculations: Dr. P. Häuselmann).
NB : Les valeurs sont normalisées pour une vie moyenne de 1.34 Ma, les incertitudes représentent des valeurs de 1 sigma (réalisation des mesures 10Be : Dr. P. Kubik, PSI/ETH, Zürich ; réalisation des mesures 26Al : Lawrence Livermore National Laboratory Californie, USA ; calculs géochronologiques : Dr. P. Häuselmann).
36The first cosmogenic nuclide dating for a high-altitude karst in the French pre-Alps contributes to modern approaches to karst chronological records of environmental change. Despite the difficulties encountered (only two reliable dates obtained for 8 samples) and the questions that remain, especially with respect to denudation rates, we obtained a significant result that confirms a Neogene age for the uppermost levels of the Mt. Granier cave system.
37The Mt. Granier and the Hauts de Chartreuse now belong to the exclusive club of Alpine karsts to have been dated using cosmogenic nuclides (Häuselmann and Granger, 2004, 2005; Häuselmann, 2007). Our results highlight the similarity in ages (around 4.3 Ma to 4.4 Ma) of the sediments in the uppermost passages of the Siebenhengste and of the Mt. Granier. It would now be interesting to obtain precise dates for the fill deposits in tier E3 of the Mt. Granier. Actual dating methods (Stock et al., 2005) allow indeed to date exactly these layered deposits, sealed by speleothems that are too old for U/Th dating. This tier E3 is thought to have formed at the Pliocene-Quaternary boundary (Hobléa 1999a), at a time when lowering of the base level accelerated. A precise dating would be a significant advance to calculate the incision rate of the surrounding valleys from the Pliocene to the present.
38Confirmation of the pre-Quaternary age of a large number of perched cave systems in the Alps would be in accordance with current ideas about the role of glaciations in Alpine speleogenesis (Hobléa, 2003; Audra et al., 2007), which is now thought to be less dominant than researchers in the 1970s and 1980s believed it to be.
39In addition to the scientific importance of producing reliable dates for conduit networks, such results also strengthen the public interest and heritage value of karst landscapes, both in general and for the Mt. Granier in particular. By providing a framework for the history of karst and landform development, accurate dates can be valuable elements in geo-education and geo-tourism activities introduced as part of sustainable development policies for environments that are often already part of conservation areas, such as the Hauts de Chartreuse Nature Reserve.
The authors would like to thank Philippe Audra (Polytech’Nice) and the Hauts de Chartreuse Nature Reserve for financing the dating process, and the staff of the laboratories where the samples were processed and analysed: Happy Ice Age Laboratory at Berne Faculty of Geology (dir. Prof. Schlüchter), Institut fûr angewandte Geologie from the Universität fûr Bodenkultur in Vienna (Prof. Dr. M. Fiebig), the Lawrence Livermore National Laboratory (California, USA). Many thanks also to the five referees for reviewing the manuscript, especially Arthur Palmer (State University of New York) and Greg Stock (Yosemite National Park) for their constructive and enriching comments. Special thanks finally to André Paillet, engineer computer graphic designer of the Edytem Laboratory.