Palaeoenvironments and cultural landscapes of the last 2000 years reconstructed from pollen and Coleopteran records in the Lower Rhône Valley, southern France

For the first time, a high-resolution pollen/Coleoptera joint analysis is performed on a late Holocene sedimentary sequence located in the Lower Rhône Valley. 14C dates validated by pollen data show that the bottom of the sequence is contemporaneous with the Greco-Roman period whereas the top is attributed to the present. This sequence yielded very rich pollen and insect assemblages, enabling a detailed reconstruction of the palaeoenvironment succession during the two last millennia around the site itself and more widely in the Arles plain. The very low pollen representation of trees and the near absence of tree-dependent Coleoptera suggest a marked deforestation of the area. The abundance of dung-beetles and nitrophytes is also in keeping with a strong grazing impact throughout the sequence. Three agricultural phases reflecting a growing level of human activities are identified. Phase 1 is contemporaneous with Celto-ligurian, Greek and Roman civilizations in Provence. At this time the forest cover was already largely destroyed, and pastoralism, cultivation of cereals, olives, vines and walnuts, was practised. Phase 2 is contemporaneous with a period spanning the Merovingian time and the Upper Middle Ages. It is characterized by increased agro-pastoral activities, probably related to the establishment of a monastic community at the Montmajour Abbey and to the settlement of farmers on the nearby Castellet hill. The major characteristic of agricultural phase 3 is the very high pollen percentages of Dipsacus fullonum or teasel, which was formerly extensively cultivated for cloth teasing. This early cultivation of an industrial plant, dated at La Calade to the twelfth century, is recorded for the first time in Provence. It may be connected with craft industries performed by monks at Montmajour Abbey. The sedimentological data suggest a succession of two stability phases interrupted by three flood phases. This interpretation agrees both with insect data, the fluctuating abundances of which are certainly connected with the alternation of ground submersion and dry periods, and with pollen data, marked by the impact of the floods upon the marshy veg etation. However, no clear climatic signal is recorded.