Abstract. We drilled a 210 m-thick succession of Quaternary sediments and extended it 30 m upsection with information that we collected from an adjacent outcrop. In the 240 m-thick succession we identified 12 different lithofacies, grouped them into five facies assemblages, and distinguished two major sedimentary sequences. A sharp contact at 103 m depth cuts off cross-beds in sequence A and separates them from the overlying horizontal beds in sequence B. Although the lowermost facies assemblage of each sequence includes a till deposited during a period of ice cover, the two tills differ from each other. In particular, the till at the base of sequence A is dominated by large clasts derived from the underlying Molasse bedrock, whereas the till at the base of sequence B has no such Molasse components. Furthermore, the till in sequence A bears evidence of glaciotectonic deformation. Both tills are overlain by thick assemblages of subaqueous, most likely glaciolacustrine and lacustrine facies elements. The cross-bedded and steeply inclined sand, gravel, and diamictic beds of sequence A are interpreted as deposits of density currents in a subaqueous ice-contact fan system within a proglacial lake. In contrast, the lacustrine sediments in sequence B are considered to record a less energetic environment where the material was most likely deposited in a prodelta setting that gradually developed into a delta plain. Towards the top, sequence B evolves into a fluvial system recorded in sequence C, when large sediment fluxes of a possibly advancing glacier resulted in a widespread cover of the region by a thick gravel unit. Feldspar luminescence dating on two samples from a sand layer at the top of sequence B provided uncorrected ages of 250.3 ± 80.2 and 251.3 ± 59.8 ka. The combination of these ages with lithostratigraphic correlations of sedimentary sequences encountered in neighboring scientific drillings suggests that sequence B was deposited between Marine Isotope Stage 8 (MIS 8; 300–243 ka) and MIS 7 (243–191 ka). This depositional age marks the end of one stage of overdeepening–fill in the perialpine Aare Valley near Bern.

The geometry of glacial overdeepenings on the Swiss Plateau close to Bern was inferred through a combination of gravity data with a 3D gravity modelling software. The target overdeepenings have depths between 155 and > 270 m and widths between 860 and 2400 m. The models show incisions characterized by U-shaped cross-sectional geometries and steep to over-steepened lateral flanks. Existing stratigraphic data reveals that the overdeepenings were formed and then filled during at least two glacial stages, which occurred during the Last Glacial Maximum (LGM) within the Marine Isotope Stage (MIS) 2, and possibly MIS 6 or before. The U-shaped cross-sectional geometries point towards glacial erosion as the main driver for the shaping of the overdeepenings. The combination of the geometries with stratigraphic data suggests that the MIS 6 (or older) glaciers deeply carved the bedrock, whereas the LGM ice sheet only widened the existing valleys but did not further deepen them. We relate this pattern to the different ice thicknesses, where a thicker MIS 6 ice was likely more powerful for wearing down the bedrock than a thinner LGM glacier. Gravity data in combination with forward modelling thus offers robust information on the development of a landscape formed through glaciers.

No abstract

Abstract. The extent and distribution of glaciers on the Swiss Plateau during the Last Glacial Maximum (LGM) can be determined from the geological record. However, similar reconstructions for the glaciations that preceded the LGM are far more difficult to be made due to the destruction of suitable sedimentary records through recurring glaciations or due to the inaccessibility of preserved records. Here, we explored Quaternary sediments that were deposited during the Marine Isotope Stage (MIS) 8 glaciation at least around 250 ka, and which were recovered in a drilling that was sunk into an overdeepened bedrock trough west of Bern (Switzerland). We analyzed the sediment bulk chemical composition of the deposits to investigate the supply of the material to the area by either the Aare Glacier, the Saane Glacier, or the Valais Glacier, and we complement this investigation with the results of heavy mineral analyses and geochemical information from detrital garnet. The potential confluence of the Valais and the Aare glaciers in the Bern area makes this location ideal for such an analysis. We determined the sediment bulk chemical signal of the various lithological units in the central Swiss Alps where the glaciers originated, which we used as endmembers for our provenance analysis. We then combined the results of this fingerprinting with the existing information on the sedimentary succession and its deposition history. This sedimentary suite is composed of two sequences, Sequence A (lower) and Sequence B (upper), both of which comprise a basal till that is overlain by lacustrine sediments. The till at the base of Sequence A was formed by the Aare Glacier. The overlying lacustrine deposits of an ice-contact lake were mainly supplied by the Aare Glacier. The basal till in Sequence B was also formed by the Aare Glacier. For the lacustrine deposits in Sequence B, the heavy mineral and garnet geochemical data indicate that the sediment was supplied by the Aare and the Saane glaciers. We use these findings for a paleogeographic reconstruction. During the time when Sequence A and the basal till in Sequence B were deposited, the Aare Glacier dominated the area. This strongly contrasts with the situation during the LGM, when the Aare Glacier was deflected by the Valais Glacier towards the northeast. The Valais Glacier was probably less extensive during MIS 8, but it was potentially present in the area, and it could have been essential for damming a lake in which the material supplied by the Aare and the Saane glaciers accumulated. In conclusion, combining provenance with sedimentological data, we could document how sediment was supplied to the investigated overdeepened basin during the MIS 8 glacial period and how glaciers were arranged in a way that was markedly different from the LGM.

<p>Pollen analysis has shown that at regional to subcontinental scales interglacial vegetation successions usually follow similar patterns throughout the Middle and Late Quaternary. Yet, relevant compositional and species-specific differences among interglacials exist and contribute to their characteristic biostratigraphic fingerprint. We use these divergences to indirectly date palaeolake deposits from Spiezberg (SW margin of overdeepened lake Thun, Switzerland), which have been assigned to either MIS 5e (last interglacial) or MIS 7a (penultimate interglacial) according to the average IRSL age of 164 ± 16 ka. For this purpose, we analyzed the fossil pollen record of the Spiezberg palaeolake deposits and applied optimal partitioning and the broken stick model for zonation. Furthermore, we re-assessed local pollen assemblage zones (LPAZ) of two physically dated reference records (Beerenmösli: MIS 5e; Meikirch: MIS 7a) from the study area by using the same zonation approaches. LPAZ of all three records were then compared optically and numerically (PCA and significance tests) to test which of the two reference records is more similar to Spiezberg. Pollen data show that the major part of the Spiezberg record was sedimented during full interglacial conditions. The irrelevance of <em>Fagus</em> points towards an MIS 5e age for the sediments. This is supported by PCA axis 2, which shows a statistically significant similarity of the Spiezberg record to the MIS 5e reference. This outcome is explained by <em>Fagus</em> playing an important role in the PCA axis 2 gradient, which is probably determined by competition for light. PCA axis 1 is not taken into consideration for the correlation since it represents a climate gradient (e.g. from boreal to temperate forests) and explains most of the intra-locality but less of the inter-locality variance. We thus assign the Spiezberg record to the last interglacial and show that climate-driven compositional differences between MIS 5e and MIS 7a are not diagnostic. In contrast, distinct differences in single-species abundances (e.g. <em>Fagus</em>) are strong enough to significantly distinguish between MIS 5e and MIS 7a records. We conclude that variability across various interglacial vegetation successions (e.g. MIS 5e vs. 7a, PCA axis 2) is smaller than local vegetation variability within a full interglacial succession at the same site (PCA axis 1). This implies a very high vegetation resilience to glacial-interglacial climatic variability, the biotic properties involved might be migration capacity, survival in refugia and adjustments to rapidly changing Quaternary environments.</p>