Grotte de Mardourneille

From the entrance to “l’Avaleur”: following the stream upstream, one reaches the foot of a pretty porch cluttered with a few blocks. One observes a concreted sill intended to measure the flow of the spring and a grille closing the porch to protect the catchment installations. One immediately enters a fine chamber 12 metres long by 5 metres wide at the start, for 4 to 5 metres in height. The floor is largely occupied by water while a sandy beach lies on the left side.

The gallery soon narrows at a small cascade and one then progresses in a straight corridor about one metre wide which stops 15 metres further on, at the threshold of the first sump. It presents itself as two joints connected by a flat-out passage (floor at -5m).

One must then climb a 4-metre cascade to regain the gallery and, after a small lake, one arrives at the second sump. It is a fine gallery 20 metres long by 2 metres high and 1.5 metres wide, with water marks. Let us specify, though you have certainly understood, that this description corresponds to exploration conditions once the sumps are empty. An ascending step and a constriction carry us into a set of small conduits that join above a fairly spacious chamber, a sort of crossroads of the network, at the level of the third sump called “l’Avaleur”.

Towards the M 3: on the left, that is to say towards the East, a spacious gallery (5 metres wide), with a sandy floor, rises then turns into a flat-out passage. After thirty metres, a passage between blocks leads to the second entrance of the network, the M 3 entrance (+16 relative to the spring).

Affluent de la Vierge Muette: it is a tributary of the main course; it represents 1/5 of the flow at low water. It begins in a recess of the chamber with a tiny sump (sump 5) which is in fact only a wet arch since its sill was lowered. At the start, follow the active stream in a small tortuous conduit which continues as a meander in which one stands up. The rest is reached through a high opening on the right.

The active stream continues for 15 metres to a small artificial dam built to bring the water directly to the exit so as not to fill “l’Avaleur” during pumping. Upstream of the dam, the active stream could be followed for about thirty metres in a very narrow meander. The main conduit is more spacious; it first rises in a puddingstone zone then crosses marls through a series of small forced conduits with a sandy floor, very pleasant compared with what follows.

One then finds the stream again in a magnificent conduit lined with scallops. Downstream, one can follow the active stream for about ten metres. Upstream, one progresses easily for thirty metres to better appreciate the rest. A succession of low passages, floor occupied by pools full of water, hard calcite for joints weakened by repeated baths (kneepads indispensable, especially for old knees) unfolds over about a hundred metres. Two more aquatic passages in which one must immerse oneself up to the chest and a delicate meander to negotiate spice up this section.

It is followed by a very narrow meander (Egyptian style) in the puddingstones but at least one can pass it standing up (well, the slimmest ones, the sturdy ones gave up long ago…) to arrive in front of the highlight of the tributary: an almost wet arch; complete immersion to pass it.

This time, it is certain, one is completely wet and the violent draught is there to remind one of it. Fortunately the rest is relatively comfortable: one follows a fine meander in the puddingstone for about a hundred metres to a last dogleg beyond which the most daring have managed to progress 7-8 metres in a very narrow ascending meander. If there are any takers, we will be pleased to accompany them.

From “l’Avaleur” to the terminal sump: on the right is sump 3 nicknamed “l’Avaleur” because it swallowed a wetsuit leg with a karabiner and a plastic crate. It is indeed filled with sand accumulated at this low point. After 20 metres towards the west, the conduit turns towards the south-west and rises then opens into a wide chamber. There, a collapse forces one back to water level and one finds the active stream again in a fine gallery 50 metres long, a little crumbling but measuring up to 10 metres high and 5 metres wide.

After the slightly aquatic passage of a pool, a new collapse bars the gallery marking the beginning of sump 4. This sump is partly due to the collapse of the gallery which caused a back-up upstream, probably of at least several hundred cubic metres, according to the measurements made during pumping. One follows the ceiling which descends, interspersed with bells, then one emerges into a horizontal part after a sandy flat-out passage. The gallery is wide, 4 to 5 metres, framed by fractures and cluttered with dunes of sand and mud. At the low point, the sand almost reaches the vault which plunges again and could not be passed.

Affluent de la Vierge Bavarde: just after the first collapse of the main active stream, a small gallery, at first hardly visible, heads north. Composed at the start of two parallel galleries, it then turns into a fine meander with a clover-shaped profile. Further on, one arrives at the “Marmite Trouée” at a clear widening.

The gallery then turns towards the west and enlarges considerably. At this point, the most practical passage is high up above a large block; one enters a keyhole meander. Above the large block, through a small hole, one reaches the bottom of a chimney climbed to + 27m.

Progression then resumes towards the north for more than 40 metres. Another chimney rises 8 metres in this sector. The meander opens into a pretty chamber at + 12m.

Two passages join higher up in a new meander where one reaches the puddingstones. A series of ascending constrictions allows one to emerge into another magnificent meander in the marls which runs south for about fifty metres. One notices there a magnificent giant’s kettle in the puddingstones. It ends at a climb of about ten metres at the base of which the marl/puddingstone contact is spectacular.

At the top, one crosses this layer of puddingstones again and reaches a spacious gallery dug in the marls which ends, 30 metres further on, at the foot of a narrow arrival in the ceiling.

The rest is a little before, towards the north. On the side and high up, a low gallery a few metres long then enlarges and penetrates the puddingstone bed. After about a hundred metres, a crossroads appears. A lower gallery of meander type winds in the puddingstone and emerges, after 80 metres, at the ceiling of the previous chamber. This section is of great beauty but a harsh route. The main conduit continues for several hundred metres, still dug in the puddingstones, which produces astonishing landscapes with many pebbles of non-limestone rocks protruding in relief on the gallery walls. A few pools and awkward crawls enliven the route.

A last narrow aquatic crawl gives access to the current terminus of this branch, consisting of a thin draughting fissure at +53m relative to the entrance of the cavity.

A little before the end, several small ceiling conduits were hastily traversed.

Let us return to the marly chamber to explore the downstream of the upper gallery. Two possibilities are offered for the route: on the right (looking north), behind blocks, one discovers a hidden passage which continues as a gallery to the top of a small 4-metre step preceding a chamber (Salle des Plongeurs) which continues for about fifteen metres towards the north to a collapse in the marls. Straight ahead, in the continuation of the small upper conduit, one can follow it in a succession of relatively cramped passages interspersed with small chambers. A more severe descending crawl follows and leads to the top of a very narrow step in the puddingstone (4 metres). One arrives in a small chamber which continues as a very low and muddy flat-out passage.

After this selective and uninviting passage, the conduit continues for about forty metres while keeping very reduced dimensions. A last very narrow portion could be surveyed for 20 metres by the thinnest among us. Stop at an even narrower area, which says it all.

Nevertheless, the survey revealed that the end of this branch is only (!!) about forty metres in plan and fifteen metres above a small cavity that we discovered at the foot of a small rocky bar, the M 5, and in which a complete dig allowed us to progress about a dozen metres.

Geomorphology: the cavity is entirely similar to many active networks dug in limestone. It stands out in this from the other cavities of the plateau which have very particular forms linked to a digging at the puddingstone/marl contact. Its role as collector and its functioning as a spring are not unrelated to this. The main forms encountered are forced conduits, magnificent meanders and galleries dug on fractures, all almost entirely lined with scallops. Spectacular kettles complete the picture.

The network seems relatively young but one can distinguish several phases of digging and more recent captures. The oldest levels are the Trou du Bolot and the Affluent de la Vierge Bavarde, perched about 50 metres above the current entrance.

Obviously, the Affluent de la Vierge Bavarde seems to have been independent of the rest of the network. It flowed from SSW to NNE from point +53 to point +32 to continue to the surface in the direction of the M 5. A first capture then fossilised the downstream part and dug the network towards the Salle des plongeurs to emerge likewise in the same sector. A second capture then occurs at the marly chamber, the active stream is captured by the main course, it changes direction completely and then flows from north to south. A third capture then takes place upstream by regressive erosion and will contribute to scouring the chamber in the marls; it is still temporarily active.

Thanks to fractures, this active stream crosses an alternation of marly and conglomeratic layers to join the collector. The establishment of the latter is certainly the most recent in the history of the cavity. Indeed, the Affluent de la Vierge Muette appears earlier than it; this active stream flows from south to north and emerges through the current entrance.

The main collector, whose outlet seems to be the M 3 entrance, or at least a clogged entrance a little lower in the valley, cut this tributary and then took over this conduit, enlarging it. One thus witnesses a migration of outlets towards the East, as the base level descends. A new capture towards the springs of the Artigue, further East, is currently under way. One could establish the following relative chronology: 1- Establishment of the active stream of the Vierge Bavarde. 2- Establishment of the active stream of the Vierge Muette. 3- Establishment of the main active stream which captures the previous one. 4- Capture of the Vierge Bavarde by the collector.

It seems obvious that the dynamics created by the drop played a positive role in the digging of the network by favouring the evacuation of the important insoluble parts present in the marls constituting the upstream of the catchment basin. If a penetration towards the surface can one day be made thanks to other pumping associated with digs, one will fall back on the typical galleries of the plateau, associated with important clay deposits which risk slowing or stopping progression.

But the substantial draughts observed in the two tributaries show that connections with the surface remain possible. For perceptive explorers to flush them out and make the wildest dreams come true: an unprecedented traverse through unique rocks…

Geology - Stratigraphy: the cavity develops in the molasses of Carcassonne, deposits dating from the Eocene (Tertiary), more precisely from the Cuisian (about 40 million years). These are formations deposited by rivers coming from the Pyrenees chain during its most intense phase of uplift. These terrains consist essentially of sandstone, marls, marly limestones, clay, conglomerates and puddingstones.

The Grotte de la Madourneille has the particularity of penetrating widely into these layers and crossing them. The lower parts of the cavity are essentially dug in variegated marls of yellow and wine-lees colour and whose carbonate content must be close to 60%. In the branch of the Vierge Muette, one crosses more sandy or frankly conglomeratic levels which are included in the marly mass and form sorts of lenses. They doubtless correspond to channel terminations that are not well individualised or to first detrital arrivals in a still strongly turbid environment.

When one climbs in the series, one encounters massive deposits of puddingstones interspersed with much less powerful marly levels, clearly visible in the landscape where they form several cliff bars (Roc de l’Aigle notably). The tributary of the Vierge Bavarde allows one to cross them, which is exceptional and unique for the moment among all the existing cavities of the plateau. The attached section clearly shows this phenomenon. The gallery crosses a first puddingstone bed, then a marly level; another spectacular puddingstone bed is then climbed (see photo) up to another marly level and the gallery then establishes itself in a third puddingstone bed which it will no longer leave for several hundred metres. At its terminus, it seems to reach a new marly level.

In total, nearly 700 metres of conduits develop in the puddingstones, which makes it a unique case among French cavities to our knowledge.

These rocks do not seem to slow karstification; the wall forms are similar to those encountered in limestones, let us simply note the many insoluble pebbles that remain in relief, some completely freed holding only by a small peduncle connected to the wall; the limestone pebbles (mostly) are sculpted with scallops or kettles as if it were a compact rock.

One can estimate at about 80% the carbonate content of these deposits which thus behave like a classic limestone.

Geology - Tectonics: detailed observation of the galleries shows the preponderance of fracturing in the digging of the cavity and especially that of submeridian fractures. This fracturing had already been highlighted by a statistical study of the then known galleries of the plateau (Spélé Aude n° 3) and confirmed by a photo-fracturation study by Jean Bienvenu (Spélé Aude n° 7).

The attached direction diagram clearly illustrates this characteristic. If the fractures are hardly visible at the surface, because masked by the alluvial cover, the underground environment allows one to observe them in situ and to verify the hypotheses put forward by geologists. The brevity of the visits granted to us by the pumping did not allow us to make precise surveys; this is a research avenue that remains to be exploited.

This rivière souterraine (underground river) is captured and supplies Mayronnes, replacing the springs of the Artigue which are insufficient in the dry season.

It has been the subject of several pumping operations by the Spéléo Corbières Minervois which has brought the total of surveyed galleries to more than 1600 metres, which makes it the most important cavity of the plateau.

In 1995, the collaboration between the cavers and the Conseil Général de l’Aude materialised in the creation of a sill and the installation of a pressure sensor intended to monitor continuously the evolution of the flow of the spring. This device was to be completed by the installation of a rain gauge at the watchtower. Unfortunately, to this day, this experiment could not be carried to its term. However, the hydrograph obtained between 06/12/1995 and 10/04/1996 shows a succession of important floods corresponding to winter and spring rains.

These floods, of the order of 1 to 2 m3/s (exit gallery almost under pressure for exceptional floods), appear for rainfall above 50 millimetres. One notes that at low water the flow is low, of the order of 2 to 3 l/s. This confirms that the Madourneille system is above all transmissive and little capacitive. The dolines of the Pla de Vidalbe constitute the upstream of the system.

A priori, the sector comprised between Coume Belle and the Pech Arnaud would be independent, the drainage being towards the spring of Rabassa and the ruisseau des Crozes; this is confirmed by the fracturing and the dye tracing of Coume Belle (see Spélé Aude n°5).

Attempt at a balance: catchment area: 1.7 km². Volume flowed: 300,000 m3/year. Mean Q: 35 m3/h. The average rainfall is around 900 millimetres on the eastern part of the plateau. For many years, the plateau has experienced a significant extension of forest cover which gradually replaces the original grassland. The exclusive planting of conifers has a notable effect on evapotranspiration which tends to increase. According to research carried out by I.N.R.A. Nancy, it can exceed 50 % of annual precipitation, which is not without consequence for the flow of springs which can disappear. We have been able to verify this point across the whole of the Corbières.

Under these conditions, despite abundant precipitation, effective infiltration is probably less than 200 mm/year. Generally speaking, the water of the Madourneille is essentially calcium bicarbonate, the greater part of the mineralisation being ensured by the HCO3- and Ca2+ ions, coming from the dissolution of the carbonate cement of the conglomerates. The Na+, Cl- and SO42- ions in modest concentration come from rainwater and the leaching of the marls. The average conductivity at 20° C, which reflects the total mineralisation, measured on the main springs of the plateau, is of the order of 480 µS/cm. The average values of 375 µS/cm (minimum of 339 on 08/12/1999 and maximum of 416 on 01/10/1999) found for the Madourneille (attached table) correspond to a developed karst with rapid circulations.

We measured the water temperature of the main springs of the plateau during the high-water period and partly at low water. Overall, one has orthothermic springs; there is a good correspondence between the average air temperature and the water temperature of the springs. Karst springs, which drain the upper parts of the plateau, generally have colder temperatures; this is clearly visible for the spring of the Madourneille with an average of 11.4° C (minimum of 10.4° C on 28/01/1995 and maximum of 12.2° C on 31/12/1994), 1.5° C lower than the expected temperature; the low-water temperatures align with the line of the average air temperature.

Toponymy: the origin of the charming and mysterious name of the Madourneille could be given by an eminent linguist, Dr Jacques Lemoine. He points out the citation “Riparia de Madornelle” (1392) which he translates as Madournelle which will then give Madourneille. This name would be a deformation of the Gaulish name “Matrona”, goddess of Springs, the “Mothers”. An origin that seems plausible since the Madourneille is the most important spring of the sector and, moreover, this side of the mountain is named “Ubac des Duzès”, which can be translated as the side of the springs, “duzès” having to be a deformation of the famous Occitan “dotz” which, as everyone knows, means spring or water conduit (more appropriate to karst phenomena) and which is very frequent in the Corbières.