Constraints on Pleistocene glaciofluvial terrace age and related soil chronosequence features from vertical 10Be profiles in the Ariège River catchment (Pyrenees, France)

Abstract

Sequences of alluvial fill-terraces record the long-term variation of fluvial transport regimes in response to Quaternary climatic changes. Through a production of increasingly precise chronologies, the broadening range of sediment dating methods in recent years has improved our understanding of correlations between alluvial dynamics and external forcing mechanisms. However, results in this field have mostly focused on periglacial and other mid-latitude settings. These may not readily apply to glaciofluvial environments, where alluvial regimes are narrowly dependent on the behavior of the adjacent ice-field. Here we constrain the ages of three glaciofluvial fill terraces based on five vertical 10Be age profiles in the Ariège River catchment, and produce a chronofunction of related contrasts in terrace tread topsoil characteristics. Each profile consists of 9 to 11 gravel samples collected over depths of 4.6 to 15m. Terrace T1 (+15m above the current riverbed) was sampled at ~4, 22 and 53km downstream from the Last Glacial Maximum terminal moraine system of the Ariège outlet glacier. The profiles of T2 (+30m) and T3 (+50m) complete the data set. Beryllium-10 age profiles date the age of terrace tread abandonment as a result of fluvial incision. Results show that the incision of T1 occurred over a protracted period during the Last Termination, beginning ca. 17.5–3.5+2.0ka. Prior to this, aggradation occurred from MIS 4 to the global Last Glacial Maximum (LGM), i.e.,under maximum ice extent conditions, and lasted a few thousand years thereafter. The protracted influx of debris after the LGM was sustained by paraglacial storage release in the catchment and promoted by strong coupling between slope and channel processes. Fluvial incision of T1 occurred during the post-LGMcold-to-warm transition after a time (supported by paleobotanical evidence) when the vegetation cover was capable of ensuring catchment-wide slope stability. The wave of incision propagated downstream between 17.5–3.5+2.0ka (proximal profile) and 13.0–0.5+3.5ka (distal profiles). The exposure-age confidence intervals for T2 and T3 are much larger than for T1 (60–145ka and 204–226ka, respectively) but the data nonetheless allow the aggradation of T2 and T3 to be attributed to MIS 6 and MIS 8, respectively. The soil chronosequence permitted by the fill-terrace age sequence indicates that the Luvic Cambisols capping T1 formed after ~13–16ka, i.e.,during the Holocene; (ii) the Haplic Luvisols capping T2 required at least ~130ka, i.e.,two interglacials; and (iii) the deeply rubified, highly eluviated Luvisols capping T3 required at least ~250ka, i.e.,at least three interglacials.