Karstification des plateaux et encaissement des vallées au cours du Néogène et du Quaternaire dans les Grands Causses méridionaux (Larzac, Blandas)

Abstract

Study of the Millau travertines correlated with the Escandorgue Plio-Pleistocene and Tarn valley volcanic formations, provides a precise and complete image of the state of the southern "Grands-Causses" landscape at the out-set of the Quaternaty. In the Larzac plateau, the study of the Grands Causses ruin-like dolomites (Larzac, Causse Noir, Fig. 2) has enabled us to reveal the crypto-corrosion rule and the major phases of their evolution (Figs. 3, 4). Crypto-karst dates back a long way. We present two complementary chronological observations: 1) The Escandorgue volcanos locally fossilize large well-differentiated kluftkarren and also clearly pre-Quaternary canyons (Fig. 6). 2) The Miocene volcanism of the northern Causses (Eglazines, Azinières) and the Neogene deposits of the Languedoc piemont are compatible with contemporaneous canyon excavation. We can thus conclude that the formation and exhumation of the major ruin-like forms of the dolomitic karst date from the Neogene. Dolomitic kluftkarrens are also a good illustration of karstic immunity. The most impor-tant karstification phase also dates back to the Neogene. It was responsible for the formation of the Causses landscape. The initiation of canyon incision is most probably Early or Middle Miocene. It can, with all probability be correlated to the Burdigalian or Serravallian period, which records the first direct deposits of the Cevennes quartz pebbles in the Languedoc Miocene sea of the Herault trough. In Southern-Escandorgue, the volcano-sedimentary formation at Bernasso (Fig. 5) infills two palaeovalleys, which have been the subject of previous studies, and presents un exceptional sequence that is important for the palaeo-environmental reconstruction and palaeogeographic evolution of the southern Causses. The existence of a 150 m pre-volcanic gorge cuiting through the Hettangian limestone is confirmed. The first detrital deposits (alluvial deposits, .slope deposits, and travertine formations) reflect an environment devoid of any volcanic structure. The first pyroclastic deposits followed by Stromholian flows piled up within the prevolcanic palaeovalley and formed a dam at 2.05 + 0.06 Ma, before the lower limit of the Olduvaï suh-chron, and should be correlated to the "Tiglian" complex. The Bernasso formations confirm the importance of the peri-Caussenard pre-Quaternary erosion. The relations between five travertine stages and fluviatile terraces of the Tarn valley have permitted a comparative evaluation of the rate of sinking of one of the big Causse valleys during the Quaternary (Fig. 7). The older travertine (Plateau de France, stage 1) is 120 m higher than the Tarn river; it belongs most probably to the Early Quaternary. The second stage (Peyre I Travertine) has been dated (230 Th/234 U) at ? 350 000 years old; results from reverse paleomagnetism suggest an older age than the Brunhes/Matuyama boundary. The site of La Roquette (Creissels) belongs to the third stage of travertine (Chenil stage); the 230 Th/234 U dating of this travertine gives a minimum age of 240 000 years, but probably closer to the limit of the method, i.e. 350 000 years. The 230 Th/234 U dating of the Chateau de Creissels travertines (stage 4) which overlie the 15 m level, provide an age of 222 000 years (+ 44 000, - 33 000) and authorizes the correlation of the stage 3 travertine (Chenil - La Roquette) with one of the Middle Quaternary Interglacial periods. In La Rouquette (stage 3) a chopping-tool fashioned from a quartz pebble (exogenous to the carbonate sediment) has been found at the paleosol. In conclusion, the rhythm of excavation of the Tarn canyon at Millau was relatively slow. The chronological facts deduced from the age of the travertines and from the Azinières volcano enable us to fix the altimetric marks for the excavation of the Tarn gorges (Fig. 7). Moreover, the alluvial 125 m level underlying the Plateau de France travertine (stage 5) could indicate the position of the at the beginning of the Quaternary at the time of the Azinières volcano (Fig. 6). Its position at the top of the structural surface of the Lias broadly eroded down to the Upper Miocene (age of the volcano) - implies an excavation of the Tarn to an altitude equal for that of its topographical position, say 750 m. That is to say, at more than about 100 m down from the lips of the gorges in the neaby Causse Noir. The Vis gorge (Fig. 9), which offers some exceptional sections for the study of the Jurassic of the Causses, does not present in spite of some remarkable exceptions (Fig. 9) evidence of its Neogene and Quaternary evolution comparable to that preserved in the Tarn valley. The most explicit vestiges of the gradual sinking of the river bed remain the incised meanders steps at various levels in the gorges, the only ones to have preserved the ancient alluvium of the river (Fig. 9). The incised Navacelles meander, "a veritable morphological success" (Fig. 10) contrary to a widely held opinion is net related to a periglacial embankment, but to a carbonate morphogenesis later than the final cold phase contemporary with the beginning of the Holocene. In fact, the long phase erosion of the concave banks was followed by a phase of travertine accumulation (Fig. 11) which, in raising the bed, was the originating factor in the incision of the meander. The age of this travertinization has been established by Carbon 14 dates (6200 and 6100 + 100 B.P.), but on the other hand the point of departure of their reincision in terraces is less known (Fig. 11). This incision of about 10 m has been stuck, up river, against the Jurassic strata of the Navacelles palaeomeander - an obstacle which the river has only now begun to notch and which it cascades over. The study of the Southern margin of the Grands Causses, bringing together the basins facing the Mediterranean and the Atlantic, is a deciding factor in the understanding of the Causses landscapes. They form two complementary "bounds" which, since the beginning of the Neogene, responded to very different geodynamic influences. The concordance of the results obtained on the perimeter of the southern Causses reinforces the demonstation and tends to validate it, in accordance with the work carried out on the Southern margin of the Massif Central (Simon-Coinçon, 1989; De Goer, 1972).