LAUSR

Abstract This paper presents the first well-dated palaeoenvironmental study from the Sarlieve marsh, a unique sedimentary record in the Limagne plain of central France, where all previous studies suffered from unreliable chronologies. We developed an accurate radiocarbon-based age-depth model and performed high-resolution multi-proxy sedimentological and geochemical analysis on a new sediment core from the heart of the sedimentary basin, improving the chronology and Holocene palaeoenvironmental interpretations. We found six potential (crypto)tephra fallouts (c. 9750, 8500, 7500, 7400, 6300, and 5800 cal yr BP) that enrich the Holocene tephrostratigraphy in Limagne, five of which were not previously documented in the sediments of the marsh. We also detected, for the first time, an array of volcanic phenomena such as degassing episodes (c. 6950 and 6050 cal yr BP), ash leaching phases (10250–9750, 8500–7400, and 5800–5100 cal yr BP), and earthquakes (c. 6800, 6600, 6050, 6100, and 1600 cal yr BP), suggesting a hitherto unsuspected period of high volcanic activity in the area between 7500 and 5800 cal yr BP. This increased activity appears to have caused a massive forcing of hydrosedimentary dynamics in the catchment in the Middle Holocene, by supplying significant volumes of fine volcanic ash to the palustrine basin, and thereby questioning the current Holocene morpho-sedimentary narrative for Limagne. The basin became a permanent freshwater lake after the Mid-Holocene climatic shift, and detrital influxes likely due to anthropogenic soil erosion appeared c. 5500 cal yr BP and grew steadily after 5000 cal yr BP, with a marked lull between 3600 and 3000 cal yr BP, perhaps because of a phase of settlement abandonments during the middle and early-late Bronze age. A series of lacustrine low-stands (c. 4750–4600, 3750–3600, 3350–3200 and 2950–2800 cal yr BP) that correlate strongly with western alpine lake records do not seem to have caused significant changes in the anthropogenic impact on soils modifying the sedimentary supply. Our results suggest that this lake could have been artificially drained c. 2550 cal yr BP, several centuries earlier than previously estimated, allowing us to hypothesize that Early Iron age societies already had substantial capacity to modify the hydraulic environment. Hydromorphic conditions developed thereafter, including complex short-lived marshy phases in late Roman times.