Morphogenesis of the Sarine canyon in the Plateau Molasse, Switzerland : new data from an archaeological site

Morphogenesis of the Sarine canyon in the Plateau Molasse, Switzerland: new data from an archaeological site The 50 to 100 metre deep canyon of the Sarine River that develops north of the prealpine front in the Molasse Plateau is classically attributed to postglacial erosion. However, the discovery of a Mesolithic in den archäologischen Schichten Photo:M. Mauvilly 186 Geographica Helvetica Jg. 63 2008/Heft 3 Fig.8:Erosion rates of the Sarine River since the Late Glacial, assuming that the canyon is a postglacial feature: a) before, and b) after the discovery of the Mesolithic site of Arconciel/La Souche Taux d’érosion de la Sarine depuis le Tardiglaciaire, en supposant que le canyon est postglaciaire: a) avant et b) après la découverte du site mésolithique d’Arconciel/La Souche Erosionsrateder Saane seit dem Spätglazial,unter derAnnahme der nacheiszeitlichen Entstehung des Canyons:a) vor und b) nach der Entdeckung des mesolithischen Fundplatzes von Arconciel/La Souche Graphics: L. Braillard archaeologicalsite (~8600 years cal. BP), locatedat the bottom of the gorge five metres above the actual river bed, has evident implications for the canyon’s morphogenesis. This new chronological datum indicates that mostof thecanyon was already formed 8600years ago, and that only five metres of Molasse have eroded since that time. This implies a dramatically high erosion rate during Late Glacial to early Holocene times 0.9 to 1.48 cm/year) and, consequently, a very low erosion rate since the Atlantic period approximately 0.06 cm/year).These new archaeological and geological findings offer exciting perspectives for Quaternary research.

Morphogenesis of the Sarine canyon in the Plateau Molasse, Switzerland: new data from an archaeological site Luc Braillard, Michel Mauvilly, Fribourg 1 Introduction From its source in the Helvetic nappes Canton of Valais, Switzerland), the Sarine River stretches over 126 kilometres, passing through the Prealps Cantons of Bern, Vaud and Fribourg) and ending in the Plateau Molasse at its confluence with the Aare River, about 20 kilometres west of the City of Bern. In its lower stretch, more precisely between the two artificial lakes of Gruyère and Schiffenen S respectively N of Fribourg), the meandering course of the river is well developed, embedded in a 50 to 100 metre deep canyon in the Miocene Molasse bedrock Fig. 1 and 2).
This geomorphological feature, which bestows upon the medieval town of Fribourg most of its particular charm, is generally attributed to postglacial erosion. Thus, erosion is said to have carved epigenetic meandering gorges ontherelativelysmooth surface left bare by the retreat of the Rhône and Sarine glaciers e.g. Van der Meer 1982;Weidmann 2005). Indeed, fossil river courses have been identified in the area southwest of Fribourg either from geophysical excavations e.g. Marescot et al. 2003) or based on the evidence of natural outcrops along the present-day Sarine course Büchi 1926;Dorthe 1962). The formation and fillings of these fossil channels still lack precise dating but are attributed to the Riss/Würm period or to older Interglacials for more details, see Fasel 2005 orWeidmann 2005).
The recent discovery of an archaeological site, located at the bottom of the gorges and attributed to the Mesolithic period Early Holocene), led to a reconsideration of the morphogenesis of the Sarine canyon.
Indeed, this relatively old human occupation is a unique chronological marker that allows a more differentiated estimate of the erosion rates of the Sarine River since the Late Glacial period. The aim of this article is to present these new archaeological findings and to discuss their implications in light of the interrelationship between humans, environment and climate.
2 The archaeological site of Arconciel/La Souche Discovered in 1998, the archaeological site of Arconciel/ La Souche is situated about six kilometres upstream of the town of Fribourg. Set in the Sarine gorges, it consists of a southwest-facing rockshelter cut out of the Molasse by the Sarine River Fig. 3), and lying five metres higher than the actual river bed Fig. 4). Thanks to a 15 metre long and 5 metre deep overhang, the rockshelter offers a good refuge from bad weather as well as ideal habitat conditions Fig. 5 more than 50 000 animal remains, 10 000 artefacts made of siliceous rocks radiolarites, quartzites, flints) which include tools e.g. scrapers, arrowheads, wedges) and several bone tools and jewels made from perforated deer teeth have already been discovered Fig. 6b, c, d).
Data gathered during the first four excavation campaigns, each one lasting amonth, allow the formulation of initial theories regarding the habitation dynamics of the shelter. A radiocarbon dating of the lowest hearth found in the stratigraphic units revealed human occupation as early as 8600 years ago Fig. 7). However, during the first three quarters of the 9th millennium cal. BP, the repetitive flood cycle of the Sarine River restricted inhabitation of the site. It is only around 8200 years cal. BP that the floods ceasedor at least decreased sufficientlyallowing Mesolithic huntergatherers to settle in the shelter. These dwellings left behind various sedimentary layers up to 50 centimetres thick that are archaeologically extremely rich.
Between 7800 and 7000 years cal. BP approximately, the shelter was still inhabited, but far less frequently.
This may be due to the collapse of part of the cliff. 3 Implications of this new archaeological data for the morphogenesis of the Sarine canyon The location of the archaeological site of Arconciel/La Souche, set at the bottom of the gorge five metres above the actual river, implies that most of the canyon was already dug at least 8600 years ago. Considering the age of the canyon postglacial) and its mean depth in this portion of the river course (~100 m), it appears that two different erosion rates were responsible for the formation of the Sarine canyon rather than a single long-term average event. Indeed, about 95 metres of sediments and Molasse were eroded between the end of the last glaciation and the Mesolithic occupation, and only five metres of Molasse has since been eroded. Taking Fig. 8). These findings question the generally accepted explanation of constant erosion by the Sarine River since the Late Glacial. In the latter case, the fluvial erosion rate would have been between 0.53 to 0.67 cm/year (~100 m of Molasse and Quaternary deposits eroded during 17 000 ±2000 years).

Discussion and perspectives
The proposed scenario for the morphogenesis of the canyon, that is a first phase of rapid incision followed by stabilisation and continued slower incision, is supported by a pedostratigraphic study conducted by Van der Meer 1982) on two fluvial terraces north of Fribourg. The soils of both terraces, located at the foot of the gorge, lie 15 metres, respectively 10 metres, above the actual valley bottom. The luvisols are well developed, and completely different from the soil on the valley floor. Based onthe content ofilluviated clay present in the B-horizon, and taking a former study by Van der Meer 1976) into account, an Older Dryas age was assigned to the high terrace and a Younger Dryas age to the low terrace. The consequence is that For the present study, the Sarine canyon is assumed to be a postglacial feature, an aspect which appears generally accepted. Therefore, except for the small portion of the river that cuts a former fossil course filled with loose material, the Sarine had to erode into the Molasse bedrock in order to form the canyon.
Concerning this point, it should however be mentioned that Van der Meer 1976) does suggest the possibility of a pre-Würmian canyon that was filled with glacial or fluvioglacial sediments during the Würm, and postglacially re-excavated. Following this hypothesis, the nature of the material to be eroded would have facilitated the re)formation of the canyon, but the rates of erosion deduced above remain unchanged.
The comparison of these archaeological and geological data offers exciting perspectives for Quaternary research. Firstly, archaeological excavation in the canyon should be intensified. Secondly, sediments that are cut by the canyon tills, fluvioglacial and glaciolacustrine deposits) or that postdate its formation calcareous tufa) should be dated OSL, 14C, U/Th) in order to improve the chronostratigraphic framework.
Thirdly, the pedostratigraphic study of terraces