Introduction
Loess–paleosol sequences (LPSs) represent valuable archives of paleoenvironmental changes in response to Quaternary climatic shifts. During glacial periods, larger riverbeds that seasonally dried up, provided dust deflation areas, while river terraces and slopes were the main sites of dust accumulation (Fig. 1). LPSs typically display alternating phases of predominantly dust accumulation and phases of predominantly soil formation, reflecting changes in regional and paleoenvironmental conditions (Pye, Reference Pye1995). Thereby, several LPSs in central Europe provide high resolution paleoenvironmental records, with distinct indicators of climatic shifts such as paleosol horizons and permafrost features (Semmel, Reference Semmel1968; Bibus and Semmel, Reference Bibus and Semmel1977; Guenther, Reference Guenther1987; Antoine et al., Reference Antoine, Rousseau, J-P and Hatté1999, Reference Andrieux, Bertran and Saito2016, Reference Antoine, Coutard, Bahain, Locht, Hérisson and Goval2021; Frechen et al., Reference Frechen, Terhorst and Rähle2007; Schirmer Reference Schirmer2012, Reference Schirmer2016; Fischer et al., Reference Fischer, Jöris, Fitzsimmons, Vinnepand, Prud’homme, Schulte and Hatté2021; Lehmkuhl et al., Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021). In contrast, LPSs in southern Europe (presently with Mediterranean climate) typically have lower chronological resolution, because their loess deposits have been relocated by slope wash processes, and their paleosol units are usually reworked by bioturbation. Frost-related features are scarce in Mediterranean LPSs (Bonifay, Reference Bonifay1965; Brunnacker, Reference Brunnacker1974; Ferraro, Reference Ferraro2009; Wacha et al., Reference Wacha, Mikulčić Pavlaković, Frechen and Crnjaković2011a, Reference Wacha, Mikulčić Pavlaković, Frechen and Crnjakovićb; Boixadera et al., Reference Boixadera, Poch, Lowick and Balasch2015; Zerboni et al., Reference Zerboni, Trombino, Frigerio, Livio, Berlusconi, Michetti, Rodnight and Spötl2015; Costantini et al., Reference Costantini, Carnicelli, Sauer, Priori, Andreetta, Kadereit and Lorenzetti2018; Wolf et al., Reference Wolf, Ryborz, Kolb, Calvo Zapata, Sanchez Vizcaino, Zöller and Faust2019; Bertran et al., Reference Bertran, Bosq, Borderie, Coussot, Coutard, Deschodt, Franc, Gardère, Liard and Wuscher2021; Lehmkuhl et al., Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021).
Figure 1. Map of aeolian deposits in middle Europe based on dataset from Bertran et al. (Reference Bertran, Bosq, Borderie, Coussot, Coutard, Deschodt, Franc, Gardère, Liard and Wuscher2021) (loess, sandy loess, coversand) and Lehmkuhl et al. (Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021) (loess and loess derivates, sandy loess, aeolian sand), including the locations of the Baix and Collias loess-paleosol sequences (LPSs) (Bosq et al., Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b; Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024), and selected NW-France LPS (including Havrincourt, Renancourt, Saint-Acheul, Saint-Illiers, Saint-Pierre-lès-Elbeuf, Saint-Sauflieu, Villiers-Adam; Antoine et al., Reference Antoine, Rousseau, J-P and Hatté1999, Reference Andrieux, Bertran and Saito2016, Reference Allard, Hughes and Woodward2021), Belgium LPS (including Harmignies, Kesselt, Rocourt, Remicourt; Frechen et al., Reference Frechen, van Vliet-Lanoë and van den Haute2001; Antoine et al., Reference Antoine, Coutard, Guerin, Deschodt, Goval, Locht and Paris2016; Haesaerts et al., Reference Haesaerts, Damblon, Gerasimenko, Spagna and Pirson2016), the Rhine Valley LPS (including Achenheim, Garzweiler, Grafenberg, Nussloch, Schwalbenberg I-II, Remagen, Schatthausen; Antoine et al., Reference Antoine, Rousseau, Zöller, Lang, Munaut, Hatté and Fontugne2001, Reference Antoine, Rousseau, Moine, Kunesch, Hatté, Lang, Tissoux and Zöller2009; Rousseau et al., Reference Rousseau, Antoine, Hatté, Lang, Zöller, Fontugne and Othman2002; Schirmer, Reference Schirmer, Ikinger and Schirmer2002, Reference Schirmer2012, Reference Schirmer2016; Frechen et al., Reference Frechen, Terhorst and Rähle2007; Gocke et al., Reference Gocke, Hambach, Eckmeier, Schwark, Zöller, Fuchs, Löscher and Wiesenberg2014; Fischer et al., Reference Fischer, Hambach, Klasen, Schulte, Zeeden, Steininger, Lehmkuhl, Gerlach and Radtke2019, Reference Fischer, Jöris, Fitzsimmons, Vinnepand, Prud’homme, Schulte and Hatté2021), the Ebro Valley LPS (including Guiamets, Tivissa, Mas de l’Alerany, Batea, Almenara; Boixadera et al., Reference Boixadera, Poch, Lowick and Balasch2015), the Po Valley LPS (including Bagaggera, Monte Netto, Val Sorda; Ferraro et al., Reference Ferraro, Terhorst, Ottner and Cremaschi2004; Ferraro, Reference Ferraro2009; Zerboni et al., Reference Zerboni, Trombino, Frigerio, Livio, Berlusconi, Michetti, Rodnight and Spötl2015; Costantini et al., Reference Costantini, Carnicelli, Sauer, Priori, Andreetta, Kadereit and Lorenzetti2018), Susak LPS (including Sand Pit, Bok, Savudrija; Wacha et al., Reference Wacha, Mikulčić Pavlaković, Frechen and Crnjaković2011a, Reference Wacha, Mikulčić Pavlaković, Frechen and Crnjakovićb, Reference Wacha, Rolf, Hambach, Frechen, Galović and Duchoslav2018; Zhang et al., Reference Zhang, Rolf, Wacha, Tsukamoto, Durn and Frechen2018). Data: Digital Elevation Model of Europe (1 km × 1 km) and shapefile of the Mediterranean biogeographic region (from the 11 biogeographic regions of Europe defined by the European Environment Agency (EEA), 2021).
LPSs in the transition zone along a N–S gradient from central to southern Europe have remained understudied. However, given the currently ongoing climate change, it is crucial to understand past environmental responses of such sensitive transition zones to climate changes, particularly in areas of expected northward shift of the boundary between Mediterranean and temperate climates (Ottlé et al., Reference Ottlé, Etchevers, Golaz, Habets, Noilhan, Martin and Ledoux2001; Gao and Giorgi, Reference Gao and Giorgi2008; Giorgi and Lionello, Reference Giorgi and Lionello2008). Most Mediterranean LPSs are disconnected from those in central Europe by high mountain ranges (e.g., the Pyrenees, Alps, Apennines, Dinaric Alps, Pindus mountains) (Fig. 1). Thus, the LPSs along the Rhône Rift Valley in SE France represent unique archives, as the N–S orientated Rhône Rift Valley is presently marked by a change from temperate to Mediterranean climate conditions with the climatic transition around 45°N, near Valence (Météo-France, 2020).
Loess along the Rhône Rift Valley was studied in the past by Suen (Reference Suen1934), Mazenot (Reference Mazenot1956), Bourdier (Reference Bourdier1958), Bonifay (Reference Bonifay1965), Dubar (Reference Dubar1979) and, more recently, by Franc et al. (Reference Franc, Moine, Fülling, Auguste, Pasty, Gadiolet, Gaertner and Robert2017) and Bosq et al. (Reference Bosq, Bertran, Degeai, Kreutzer, Queffelec, Moine and Morin2018, Reference Bosq, Bertran, Degeai, Queffelec and Moine2020a, Reference Bosq, Bertran, Degeai, Queffelec and Moineb). According to loess maps by Haase et al. (Reference Haase, Fink, Haase, Ruske, Pécsi, Richter, Altermann and Jäger2007) and Lehmkuhl et al. (Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021), the distribution of aeolian sediments is patchy in the southern Rhône Rift Valley, whereas the map of Bertran et al. (Reference Bertran, Bosq, Borderie, Coussot, Coutard, Deschodt, Franc, Gardère, Liard and Wuscher2021), derived from particle size distributions of topsoils (thus, also including shallow completely pedogenized loess deposits), suggests a more widespread distribution of loess and sandy loess. Furthermore, the map of Bertran et al. (Reference Bertran, Bosq, Borderie, Coussot, Coutard, Deschodt, Franc, Gardère, Liard and Wuscher2021) and observations by Bosq et al. (Reference Bosq, Bertran, Degeai, Kreutzer, Queffelec, Moine and Morin2018, Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b) suggest a southward change from loess to sandy loess, whereby the region near Valence (∼45°N) again represents a transition area. The Baix LPS is located within that transition area, 25 km southwest of Valence (Fig. 1).
A preliminary chronostratigraphic framework and field observations on the lower part of the Baix LPS have been introduced previously (Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024). Here, we present an improved chronology for the entire LPS, as well as soil-color measurements, particle size analyses, chemical and micromorphological analyses, and we derive a reconstruction of the paleoenvironmental changes at the site over the Late Pleistocene. The improved chronology includes ages by accelerator mass spectrometry (AMS) radiocarbon dating on gastropod shells from the upper part of the LPS and optically stimulated luminescence (OSL) dating on coarse feldspar grains, polymineral fine grains, and quartz fine grains from six sediment blocks (OSL blocks). The results of these analyses are then used to answer our main research questions: (1) which sedimentation dynamics can be inferred from granulometric characteristics, (2) which paleoenvironmental interpretation can be derived from the geochemical characteristics of Baix LPS, (3) in what chronological frame the inferred paleoenvironmental took place, and (4) how Baix LPS compares to other European LPSs.
We hypothesize that during the last interglacial period (Eemian, MIS 5e), the Mediterranean conditions extended farther north, compared to the present boundary of Mediterranean conditions. The location of the Baix LPS is ideal to test this hypothesis, because it is located at the present northern boundary of Mediterranean conditions, which means that its paleosol of the last interglacial period should be a Mediterranean paleosol, if the hypothesis is true. Furthermore, we test how the texture of the Baix LPS fits into the earlier postulated change from loess to sandy loess in the Valence area (Bosq et al., Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b; Bertran et al., Reference Bertran, Bosq, Borderie, Coussot, Coutard, Deschodt, Franc, Gardère, Liard and Wuscher2021).
Regional setting
The Baix LPS (44°42′36.30″N, 4243′20.80″E, 155 m asl) is located in a former loess pit, 25 km southwest of Valence, at the western flank of the Rhône Rift Valley (Fig. 1). Bounded by the Massif Central to the west and the Alps to the east, the Rhône Rift Valley is mainly filled with Oligocene to Pleistocene deposits (BRGM, 2021). It is characterized by an alternation of wide basins and narrow passages, interspersed with several smaller side basins.
According to its geochemistry, loess in the Rhône Rift Valley originated mainly from glacial and fluvioglacial deposits from the Alps, with only minor contribution from other local alluvium (Bosq et al., Reference Bosq, Bertran, Degeai, Queffelec and Moine2020a). The loess mainly covers the plateaus, large basins, and terraces of the Rhône River and its tributaries (Suen, Reference Suen1934; Mazenot, Reference Mazenot1956; Bosq et al., Reference Bosq, Bertran, Degeai, Kreutzer, Queffelec, Moine and Morin2018). The valley is drained by the Rhône River, to which the eastern tributaries of the Rhône River (e.g., Isère, Drôme, Durance), originating in the Alps, deliver more sediment to the Rhône River than do the shorter tributaries (e.g., Ardèche, Gardon, Cèze) from the Massif Central (Olivier et al., Reference Olivier, Amoros, Carrel, Malard, Lamouroux, Bravard, Tockner, Uehlinger and Robinson2009).
The present-day climate in the Rhône Rift Valley is characterized by a gradual change from temperate climate conditions in the north to Mediterranean conditions in the south. The temperate climate of the northern rift valley is characterized by warm summers (up to three dry months) and cold, humid winters. The mean annual temperature (MAT) is ∼13°C, mean annual precipitation (MAP) ranges between 800 mm and 900 mm (Météo-France, 2025a, b). The mountainous areas and regions to the west of the valley reach higher MAP of 1200–1400 mm and summer storms are common (Diodato et al., Reference Diodato, Bellocchi, Romano and Guadagno2016).
The Mediterranean climate in the southern part of the rift valley is characterized by hot, dry summers with five arid months and humid winters, MAT is ≥ 14°C and MAP is ≥ 700 mm (Météo-France, 2025c, d). The wind system of the valley is influenced by advective air masses of different origins (Mediterranean, maritime, polar maritime, and subtropical), which lead to high yearly and seasonal variability of precipitation and wind intensities (Diodato et al., Reference Diodato, Bellocchi, Romano and Guadagno2016). The dominating wind system is the Mistral, blowing from the northwest through the Rhône Rift Valley, which acts as a channel, accelerating the wind, before opening into the Mediterranean Sea (Obermann et al., Reference Obermann, Bastin, Belamari, Conte, Gaertner, Li and Ahrens2018).
The Baix LPS is situated in the wind shadow, behind a small hill at the edge of an opening in the western shoulder of the Rhône Rift Valley, where a small side basin, filled with debris of Cretaceous limestone and basalt feeds into the Rhône Rift Valley (Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024). Its topographic position explains the occurrence of a loess patch just at the Baix LPS. The source of the loess at the Baix LPS was the temporarily dried-out riverbed of the Rhône River, as has been proven for other loess sites along the Rhône Rift Valley (Bosq et al., Reference Bosq, Bertran, Degeai, Queffelec and Moine2020a). This does not exclude some possible component of material from more proximal sources such as riverbeds of smaller local rivers (e.g., Payre, l’Ouvèze).
Material and methods
Sampling and profile description
We cleaned and described the main profile “Baix” with a total thickness of 13.7 m (Fig. 2) during three field campaigns between 2018 and 2020 (Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024). An additional profile, “Baix_west” (thickness 8.5 m; Fig. 2), was cleaned to assess the spatial variability of the upper sediments and paleosol horizons.
Figure 2. (A) Profile sketch of main profile Baix and “Baix_west” with soil horizon designations. Green boxes mark sample positions for optically stimulated luminescence (OSL) dating, black rectangles mark positions of thin-section samples. (B) Profile photo of main profile Baix (Photos: D. Sauer, 2017–2020).
The main profile was sampled in 10-cm increments for analyzing chemical composition and particle size distribution. Where horizon boundaries occurred in less than 10-cm distances, the sampling increments were decreased accordingly, to avoid sampling across horizon boundaries. Baix_west was sampled by taking mixed samples per horizon and horizons thicker than 30 cm were subdivided. The samples were air-dried, sieved (< 2-mm mesh size) and homogenized. Four undisturbed samples for micromorphological analysis were taken from the paleosol horizons Bw1, 5 Bw2, 6 Bw3, and 7 Btg. For this purpose, Kubiëna boxes were carved into the profile wall by a knife (Fig. 2).
Ten samples of bulk sediment, each of approximately 10 L, were taken at regular intervals from the upper 7 m of the main profile to collect mollusc shells for 14C-AMS dating at the Laboratoire des Sciences du Climat et de l’Environnement (LSCE) de Gif-sur-Yvette (France). The samples were wet sieved through a 425-mm mesh, and the mollusc shells were cleaned by ultrasonic treatment.
Six block samples for conventional OSL dating at the Heidelberg Luminescence Laboratory were taken. Three samples (OSL 12020, OSL 22020, OSL 32020,) from the 7 Bt1, 6 BCk/Btg, and 5 BCk5 horizons of the main profile Baix (Figs. 2, 6b) were collected. The other three samples (OSL 62018, OSL 72018, OSL 82018) were collected from the 5 Bw2, 3 BCk3, and 2 BCk2 horizons from profile Baix_west (Figs. 2, 6a, 6b).
Paleosol horizons were described according to the FAO guidelines for Soil Description (FAO, 2006), and the paleosols were afterwards tentatively classified according to the World Reference Base for Soil Resources (IUSS Working Group WRB, 2022). Horizon designations, detailed descriptions of the field characteristics, and stratigraphic sequences of the main profile Baix are in Pfaffner et al. (Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024, Table 1).
Based on the preliminary results in Pfaffner et al. (Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024), we distinguished from bottom to top: a last interglacial to early glacial paleosol complex (IG-EP-PSC), an early pleniglacial paleosol (EPG-PS), a middle pleniglacial paleosol complex (MPG-PSC), and the Holocene soil (Figs. 3, 6a, 6b). These paleosol complexes are intercalated by units of unweathered or only slightly altered, yet sometimes reworked (by slope wash), loess deposits with varying degrees of weathering.
Figure 3. Stratigraphic sections of the main Baix loess–paleosol sequence (LPS) and Baix_west LPS, Rhône Rift Valley, SE France. Gray horizontal bars mark positions of paleosols. Depth curves of geochemical data (CaCO3, SOC), ratios (CPA, WIMER, Si/Al, Fed/Fet, Mnd/Mnt), and color parameters (L, a, b, a/b) of both profiles. IG-EG-PSC = last interglacial to last early glacial paleosol complex; EPG-PS = last early pleniglacial paleosol; MPG-PSC = middle pleniglacial paleosol complex; Holocene-S = Holocene soil; ED = erosional discontinuity.
By correlating the previously designated horizons and stratigraphic sequences of the main profile (Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024) with the profile Baix_west, small differences in the positions of the upper horizons suggest a gentle southwestern inclination of the sediment layers, which matches with the local relief. However, the main trends of the weathering proxies of Baix_west match well with those of the upper part of the main profile down to the MPG-PSC at 8.5 m below ground level (Fig. 3). Because of this great similarity, we focus on the results of the main profile Baix.
Laboratory analyses
Particle size distribution and related proxies
High-resolution particle size distribution (PSD) data provide valuable information about the transport distance of aeolian sediments under consideration of the respective environmental factors such as topography, microclimate, vegetation, and influence of post-depositional processes (Pye, Reference Pye1995; Rousseau et al., Reference Rousseau, Antoine, Hatté, Lang, Zöller, Fontugne and Othman2002; Vandenberghe, Reference Vandenberghe2013; Lehmkuhl et al., Reference Lehmkuhl, Zens, Krauß, Schulte and Kels2016; Újvári et al., Reference Újvári, Kok, Varga and Kovács2016a; Schulte et al., Reference Schulte, Sprafke, Rodrigues and Fitzsimmons2018; Vandenberghe et al., Reference Vandenberghe, Sun, Wang, Abels and Liu2018; Varga et al., Reference Varga, Újvári and Kovács2019). PSD was measured with a laser diffraction particle size analyzer (Beckman Coulter LS 13320), from samples that were prepared following ISO 11277 (International Standards Organization, 2020a) and 13320 (International Standards Organization, 2020b). For PSD determination, an optical model according to the Mie theory with RI 1.55i-0.01 suitable for quartz was applied (International Standards Organization, 2020a, b). The statistical parameters, such as mean and mode of particle size of the 116 particle size classes from 0.04 µm to 2000 µm were calculated. The percentage PSD and their shifts over depth were visualized in heatmaps. For reconstructing paleoenvironmental conditions, the GSI (26–52 µm/<26 µm) and the U ratio (16–44 µm/5.5–16 µm) were used, each influenced by different factors (e.g., wind speed, distal and local transport, post-depositional processes) (Vandenberghe et al., Reference Vandenberghe, Mücher, Roebroeks, Gemke, van Kolfschoten and Roebroeks1985; Antoine et al., Reference Antoine, Rousseau, Moine, Kunesch, Hatté, Lang, Tissoux and Zöller2009: Újvári et al., Reference Újvári, Kok, Varga and Kovács2016a; Schulte et al., Reference Schulte, Sprafke, Rodrigues and Fitzsimmons2018; Schulte and Lehmkuhl, Reference Schulte and Lehmkuhl2018). Additionally, the fine to coarse clay (< 0.04 µm/0.63–2.0 µm) ratio was applied as proxy for illuvial clay paleosol horizons.
Soil chemical analysis, weathering indices and soil color
Total contents of Fe, Al, Ca, Mg, Si, and Mn were obtained by mixing the samples with lithium meta-tetraborate (ratio 1:4), fusing this mixture, dissolving the fused samples with 5% HNO3 and analyzing them by inductively coupled plasma optical emission spectrometry (ICP-OES, iCAP 7000, Thermo Scientific). The total contents of Na and K were determined by flame spectrometer (Elex 6361, Eppendorf). Major element concentrations were used to calculate geochemical indices based on their molar ratios to identify the loss of bases from silicate weathering and subsequent element leaching due to pedogenesis. The following indices were used: the WIMER (Weathering Index based on Molar Element Ratios) calculated as WIMER = (Ca (after subtraction of Ca in CaCO3) + Mg + K + Na)/Al (Sauer, Reference Sauer, Richardson, Castree, Goodchild, Kobayashi, Liu and Marston2016) and the CPA index (Chemical Proxy of Alteration), calculated as CPA = Al2O3/(Al2O3 + Na2O)*100 (Sheldon and Tabor, Reference Sheldon and Tabor2009; Buggle et al., Reference Buggle, Glaser, Hambach, Gerasimenko and Marković2011). The Si/Al ratio was applied as a proxy of the aluminosilicates/quartz ratio (Liang et al., Reference Liang, Sun, Beets, Prins, Wu and Vandenberghe2013). Fe, Al, and Mn contents in the form of pedogenic oxides/hydroxides (Fed, Ald, and Mnd) were extracted by sodium dithionite Na2S2O4 (Schwertmann, Reference Schwertmann1964; Sparks et al., Reference Sparks, Page, Helmke and Loeppert1996) and analyzed by ICP-OES (iCAP 7000, Thermo Scientific). Fed/Fet ratio and Mnd/Mnt ratio as depth functions were used to reflect the degree of weathering and hydromorphic influence (Torrent and Nettleton, Reference Torrent and Nettleton1979; Arduino, Reference Arduino1984; McKenzie, Reference McKenzie, Dixon, Weed and Dinauer1989; Durn et al., Reference Durn, Slovenec and Čović2001). Furthermore, the formation of manganese oxides (Mnd) is related to seasonal water logging in soils, thus reflects the degree of intensity of paleo-redox processes (McKenzie, Reference McKenzie, Dixon, Weed and Dinauer1989).
Carbonate contents were measured by the gas-volumetric method using a Scheibler apparatus and calculated as calcium carbonate equivalents in wt% (Blume et al., Reference Blume, Stahr and Leinweber2011). Soil organic carbon (SOC) contents were estimated by quantifying total carbon (C) contents by use of a CHN elemental analyzer (TruSpec Micro, LECO) in wt%, and subtracting the CaCO3 carbon contents from the total C contents. This approach only allows for obtaining trends, while an exact calculation of SOC contents is not possible, given the error of the Scheibler method and the high CaCO3 contents of the loess in this region.
Color changes with depth in LPS reflect varying degree of pedogenesis and changing paleoenvironmental conditions (Pécsi and Richter, Reference Pécsi and Richter1996; Sprafke, Reference Sprafke2015; Wang et al., Reference Wang, Song, Zhao and Li2016). Quantitative color determination has received more attention recently (e.g., in LPS studies of Chen et al., Reference Chen, Ji, Balsam, Chen, Liu and An2002; Gocke et al., Reference Gocke, Hambach, Eckmeier, Schwark, Zöller, Fuchs, Löscher and Wiesenberg2014; Sprafke, Reference Sprafke2015; Wang et al., Reference Wang, Song, Zhao and Li2016; Zeeden et al., Reference Zeeden, Krauß, Kels and Lehmkuhl2017; Sprafke et al., Reference Sprafke, Schulte, Meyer-Heintze, Händel, Einwögerer, Simon, Peticzka, Schäfer, Lehmkuhl and Terhorst2020; Krauss et al., Reference Krauss, Klasen, Schulte and Lehmkuhl2021; Laag et al., Reference Laag, Lagroix, Kreutzer, Chapkanski, Zeeden and Guyodo2023). We used a handheld spectrophotometer CM-700d (Konica Minolta) for the color measurements. Fine-grained earth samples were measured three times (10° standard observer, SCI) and the mean results were converted into the CIE color index, presenting the colors as L*a*b* values by SpectraMagic NX software (CIE, 1976). L (metric lightness function) represents the brightness or luminance (range 0–100); a and b represent the chromaticity coordinates opponent red–green scales (a) and opponent blue–yellow scales (b) (Viscarra Rossel et al., Reference Viscarra Rossel, Minasny, Roudier and McBratney2006). Although it is not possible to separate between in-situ and reworked soil material or secondary carbonate enrichments based on color values, the chromatic parameters and derived ratios are valuable tools to detect units that have been altered by pedogenic processes, as already minor decarbonatation, accumulation of soil organic matter, or iron oxide formation will lead to measurable changes of these values.
For thin section preparation the pore water of the undisturbed, orientated blocks was gradually exchanged with acetone. The blocks then were impregnated with resin (Palatal), hardened over six weeks, cut into slices that were then cut into smaller blocks, which were glued on glass slide (28 mm × 48 mm) and ground to < 25-μm thickness (first by machine, in the end by hand). The thin sections were described according to Stoops (Reference Stoops2021) and microphotographs were taken with a Keyence VHX-7100/-E100 digital microscope in plane polarized and cross polarized light.
Chronostratigraphic framework
AMS radiocarbon dating
Two small mollusc taxa (Pupilla muscorum and Trochulus hispidus) were selected for dating because of their abundance in the samples. These species live on vegetation and are assumed to be suitable for chronological investigations (e.g., Pigati et al., Reference Pigati, Rech and Nekola2010; Újvári et al., Reference Újvári, Molnár and Páll-Gergely2016b). The conversion mode of the sample to CO2 and, if required, to C_graphite, the introduction mode BCA for the carbonate line and GG for the graphitization line followed Tisnérat-Laborde et al. (Reference Tisnérat-Laborde, Poupeau, Tannau and Paterne2001, Reference Tisnérat-Laborde, Thil, Synal, Cersoy, Hatté, Gauthier and Massault2015), Hatté et al. (Reference Hatté, Arnold, Dapoigny, Daux, Delibrias, Du and Fontugne2024) and Thil et al. (Reference Thil, Tisnérat-Laborde, Hatté, Kader, Noury, Paterne, Phouybanhdyt and Wacker2024). Ages (BP) derive from F14C and are reported following the recommendations of Stuiver and Polach (Reference Stuiver and Polach1977), particularly rounding off. Radiocarbon ages were calibrated to calendar ages with the IntCal20 calibration curve (Reimer et al., Reference Reimer, Austin, Bard, Bayliss, Blackwell, Bronk Ramsey and Butzin2020) using the software Calib version 8.2 (Stuiver et al., Reference Stuiver, Reimer and Reimer2021) (Supplement 1 – Table S1.1). Calibrated ages are reported as age ranges with a 95.4% confidence level.
Luminescence dating
For the present study, we subjected the six block samples to optically stimulated luminescence (OSL; Huntley et al., Reference Huntley, Godfrey-Smith and Thewalt1985) dating (HDS-1776 to HDS-1778 = OSL 62018 to OSL 82018 and HDS-1799 to HDS-1801 = OSL 12020 to OSL 32020). Fine grains (4–11 µm) and coarse grains (125–212 µm) were available, but quartz coarse grains were too scarce for luminescence dating (details of the sample preparation are given in Fig. S2.2.1 in Supplement 2). The single aliquot regeneration (SAR) approach was applied (Murray and Wintle, Reference Murray and Wintle2000). While quartz fine grains were measured with a blue-light stimulated (BLSL) SAR protocol (Murray and Wintle, Reference Murray and Wintle2000) detecting the ultra-violet luminescence emission around 340 nm, feldspar coarse grains and polymineral fine grains were measured with a post-infrared (at 60°C) infrared (at 225°C) stimulation (pIR60IR225) protocol (Thomsen et al., Reference Thomsen, Murray, Jain and Bøtter-Jensen2008) detecting the blue-violet emission around 410 nm. Details on measurement configurations, pretests for adapting the protocols to the Baix samples, parameters of data analyses, and results of the equivalent doses (De) of individual aliquots and the De distribution of a sample are given in Supplement 2 (S2.3–S2.5). For the quartz fine grains, we also provide the results of a series of tests in Supplement 2 (S2.4), which were conducted to investigate possible reasons for the unexpectedly high quartz fine grain ages. The analyses of the feldspar coarse grain measurements were accomplished by tests investigating the suitability of the Tx/Tn sensitivity curves as a proxy of (1) material source or (2) poor bleaching (Chen et al., Reference Chen, Li and Li2013; S2.3.9–S2.3.10). For dose rate estimation (S2.6), the contents of radionuclides were determined with the µDose system in the luminescence laboratory at Gießen University (Tudyka et al., Reference Tudyka, Miłosz, Adamiec, Bluszcz, Poręba, Paszkowski and Kolarczyk2018; Kolb et al., Reference Kolb, Tudyka, Kadereit, Lomax, Poręba, Zander, Zipf and Fuchs2022).
Results
Particle size distribution and related sedimentological indices
The heatmap of Fig. 4A showed highest contents of coarse silt and the lowest clay contents for the unweathered to slightly weathered units, whereas the paleosol horizons were marked by considerably higher clay contents. The Δ GSD (Fig. 4B) displayed the main mode in the coarse silt fraction (∼44–50 µm), thus a PSD typical for aeolian sediments and two subordinate maxima in the fine to medium silt fraction and in the fine to medium clay fraction. Throughout the profile, the mean particle size ranged between 29 µm and 65 µm with an average of 36 µm, while the median particle size ranged from 23 µm to 41 µm, and the average median was 30 µm (Fig. 4A).
Figure 4. Stratigraphy and grain size (GS) data of main profile Baix. (A) Heatmap depiction of grain size frequency [vol%], vertical curves present GS variations and ratios GSI (26–52 µm/<26 µm), U ratio (16–44 µm/5.5–16 µm), fine clay/coarse clay (< 0.2 µm/0.63–2 µm). (B) Heatmap of Δ grain size distribution [vol%]. fC = fine clay (0.04–0.2 µm), mC = medium clay (0.2–0.63 µm), cC = coarse clay (0.63–2 µm), fSi = fine silt (2–6.3 µm), mSi = medium silt (6.3–20 µm), cSi = coarse silt (20–63 µm), fS = fine sand (63–200 µm), mS = medium sand (200–630 µm), cS = coarse sand (630–2000 µm). IG-EG-PSC = last interglacial to last early glacial paleosol complex; EPG-PS = last early pleniglacial paleosol; MPG-PSC = middle pleniglacial paleosol complex; Holocene-S = Holocene soil; ED = erosional discontinuity.
In the field, no laminations indicative of slope wash were observed. However, most of the horizons contained minor amounts of small rock fragments (Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024, Fig. 11) suggesting gentle erosion and subsequent re-accumulation of sediment at the foot slope position where the Baix LPS is located. Apart from this observation, all paleosol horizons consisted of homogenous silts. In contrast, the intercalated unweathered and slightly weathered loess horizons showed an increase of the GSI towards the top, accompanied by a constant decrease of the silt fraction and an increase of the fine sand fraction. This trend is also reflected by higher mean and median PSD values (Fig. 4A). The U-ratio (1.55–3.03), GSI (0.34–0.79), and Si/Al (4.55–6.64) showed maxima within the unweathered and slightly weathered units and decreased in the paleosol horizons, whereas fine clay/coarse clay (0.16–0.57) ratios acting oppositely (Figs. 3, 4A).
Colors and chemical characteristics
Generally, the paleosol horizons showed lower L values than the unweathered to slightly weathered units (Fig. 3). Above the MPG-PSC, the L values followed a steady upward increase, similar to the upward increase in carbonate content and WIMER. The paleosol horizons also exhibited maximum a and b values, with a = 11 and b = 32, attributed to the pedogenic iron oxide formation, while the unweathered to slightly weathered units showed much lower a and b values, with minima of a = 5 and b = 22. The lowest a and b values coincided with the least weathered units (e.g., the Ck horizons).
Carbonate contents showed maxima around 23 wt% CaCO3 equivalent, which were reached in the Bk1 horizon of the Holocene soil and 5 BCk4 horizon of MPG-PSC (Fig. 3). In addition to these carbonate contents of the fine earth, hard carbonate nodules of various sizes were observed in most parts of the profile, from the Bk1 horizon down to the 7 Btg horizon, indicating steady carbonate dissolution and reprecipitation during the accretion of the loess. The MPG-PSC displayed considerable carbonate leaching, with carbonate contents as low as 0.3–2.5 wt% CaCO3 equivalent. The soil organic carbon (SOC) contents fluctuated between 0 and 0.5 wt%. As expected, the maximal values are highest in the Holocene soil and decrease with depth, except for the 5 Ck2 horizon (0.5 wt%).
The weathering indices of the least-weathered units exhibited minimum values for the CPA (68) and Fed/Fet (0.45) and Mnd/Mnt (0.86) ratios, and maximum values for the WIMER (0.95) and Si/Al (6.6) ratio. These weathering indices all highlighted the most advanced pedogenesis of the IG-EP-PSC, showing the maximum values for CPA (80) and Fed/Fet (0.90) Mnd/Mnt (1.3), and minimum values for WIMER (0.49) and Si/Al (4.5). The second-most advanced pedogenesis of the MPG-PSC was also expressed by a high CPA (up to 74), slightly increased Fed/Fet ratio (up to 0.60) and Mnd/Mnt ratio (up to 1.0), and low WIMER (down to 0.70) and Si/Al ratio (down to 5.0) (Fig. 3). The WIMER also reflected the pedogenesis of the Holocene soil, which was less pronounced in the other weathering indices.
Micromorphological characteristics of the paleosol horizons
The mineralogical composition of the matrix of the four investigated soil/paleosol horizons was dominated by quartz, feldspar, biotite, and calcite (Fig. 5a–d). The void system was characterized by predominance of biogenic pores in all studied horizons, mainly by channels, chambers, and some vugs. Passage and bioturbation features, and calcite earthworm granules were observed in all studied soil/paleosol horizons (Fig. 5a, e, f). Illuvial clay and clay–silt coatings occurred along the voids and on aggregate surfaces in all studied horizons (Fig. 5g, i, k, s). Primary carbonates are absent in all paleosol horizons, however secondary carbonates are present in the form of micrite hypocoatings (Fig. 5p). In all studied horizons, orthic and disorthic Fe/Mn nodules occurred (Fig. 5k, m). The 6 Bw3 and 7 Btg horizons showed two large reddish rounded aggregates containing iron oxides and fragments of limpid clay coatings (Fig. 5c, d). The broken clay coatings and the varying degree of rounding indicating that the aggregates had been relocated or reworked to varying degrees (5o). The brown paleosol Bw horizons displayed a stippled-speckled to poro-striated b-fabric (Fig. 5j, l, n), whereas the 7 Btg horizon showed a poro-striated b-fabric (Fig. 5r, t).
Figure 5. Photomicrographs of paleosol horizons and paleosol-complex of profiles Baix (Bw1: 43–51 cm; 6 Bw3: 1133–1141 cm; 7 Btg: 1209–1217 cm) and Baix_west (5 Bw2: 734–742 cm). (a–d) Photomicrographs of whole thin sections (3 cm × 4 cm). (e–t) Photomicrographs of certain details of the horizons at various magnification. Feature relations are discussed in text section “Micromorphological characteristics of the paleosol horizons”. PPL = plain polarized light; XPL = crossed polarized light; ci = clay infilling; cc = clay coating; ch = calcite hypocoating; ci = calcite infilling; cs = clay-silt coating; dz = depletion zone; eg = earthworm granule; fcc = fragmented clay coating; Fe-c = Fe coating; Fe-n = Fe/Mn nodule; pf = passage feature; ra = rounded aggregate; rr = root residue; wr = weathered rock fragments. (a) Bw1 (Baix) Subangular-blocky microstructure and channels, partially filled with rounded, bioturbated aggregates (PPL) (b) 5 Bw2 (Baix_west) Subangular-blocky microstructure and channels, partially filled with rounded, bioturbated aggregates (PPL) (c) 6 Bw3 (Baix) Massive to channel microstructure, channels partially filled with porous microaggregates, iron depletion zone; orthic and disorthic, rounded and digitated Fe/Mn nodules (Ø 350 µm) (PPL) (d) 7 Btg (Baix) Weakly subangular to angular blocky microstructure with channels and vugs; large, red, rounded anorthic nodule with Fe and clay coatings and fragmented clay coatings; common small round, digitated, elongated, disorthic Fe/Mn nodules; larger typic orthic Fe nodule (PPL) (e) Bw1 (Baix) Large passage feature in upper left corner of Figure 5a (PPL) (f) Bw1 (Baix) Passage feature showing crystallic b-fabric with calcite earthworm granule (XPL) (g) Bw1 (Baix) Calcite fillings (needle-fiber calcite) and clay–silt coatings along voids (PPL) (h) Bw1 (Baix) Same area as Figure 5g with stippled speckled b-fabric (XPL) (i) 5 Bw2 (Baix_west) Round, brown aggregate with clay–silt coatings along the edges of the aggregate (PPL) (j) 5 Bw2 (Baix_west) Same area as Figure 5i with striated and stippled speckled b-fabric (XPL) (k) 6 Bw3 (Baix) In-situ dusty clay–silt coatings at the bottom of the void with abundant anorthic Fe nodules (PPL) (l) 6 Bw3 (Baix) Same area as Figure 5k with stippled speckled b-fabric (XPL) (m) 6 Bw3 (Baix) Dusty clay–silt coatings along voids; typic to concentric orthic and anorthic Fe nodules (PPL) (n) 6 Bw3 (Baix) Same area as (m) with stippled speckled b-fabric (XPL) (o) 6 Bw3 (Baix) Large rounded, reddish anorthic nodule (Ø 2200 µm) containing Fe and clay coatings (PPL) (p) 7 Btg (Baix) Poro-striated b-fabric; calcite hypocoatings (XPL) (q) 7 Btg (Baix) Iron oxide coatings and infillings; clay coatings along voids (PPL) (r) 7 Btg (Baix) Poro-striated b-fabric; limpid clay coatings (XPL) (s) 7 Btg (Baix) In-situ limpid orange clay and gray-brown clay–silt coatings along voids (PPL) (t) 7 Btg (Baix) Same area as Figure 5s with poro-striated b-fabric (XPL).
In detail, the Bw1 horizon in the main profile Baix (43–51 cm) was characterized by a subangular blocky structure and abundant channels, few vugs, micrite hypocoatings along voids, dense calcite infillings, calcite earthworm granules (Ø 670 µm), root residues (Fig. 5a), and a crystallitic and speckled b-fabric (Fig. 5f, h). A large, round passage feature was present in the upper left corner of the thin section (Fig. 5a), including a calcite earthworm granule (Fig. 5e, f). Also, infillings of needle–fiber calcite and clay–silt coatings along voids were observed (Fig. 5g, h).
The 5 Bw2 horizon in the profile Baix_west (734–742 cm) exhibited a massive to weakly subangular blocky microstructure, many channels that were partially infilled with rounded aggregates, few vugs, abundant disorthic, rounded, elongated, digitated Fe/Mn nodules, and typic, digitated orthic Fe nodules (Ø 580 µm) (Fig. 5b). Moreover, rounded compound micro-aggregates with clay–silt coatings along the edges of the aggregates were observed (Fig. 5i, j).
The 6 Bw3 horizon in the main profile Baix (1133–1141 cm) was characterized by a massive to channel microstructure, whereby the channels were partially filled with porous microaggregates. Orthic and disorthic, rounded and digitated Fe/Mn nodules (Ø 350 µm), and a depletion zone were observed (Fig. 5c). In-situ, dusty clay–silt coatings with abundant anorthic Fe nodules occurred along voids (Fig. 5k–o).
The 7 Btg horizon in the main profile Baix (1209–1217 cm) had a weakly subangular to angular blocky microstructure with abundant channels, large, red, rounded anorthic nodules with Fe and clay coatings, fragmented clay coatings, and common small, digitate, round to elongated, disorthic Fe/Mn nodules and larger typic orthic Fe nodules (Fig. 5d). It showed a poro-striated b-fabric, calcite hypocoatings (Fig. 5p, r, t), iron oxide coatings and infillings (Fig. 5q, r), and limpid orange clay and gray-brown clay–silt coatings along voids (Fig. 5s, t). Both, in-situ and reworked limpid, layered illuvial clay coatings were observed in the 7 Btg horizon (Fig. 5d, q, s). The layered clay coatings along the void walls started with limpid orange clay coatings, followed by dusty, grayish-brown clay–silt coatings, and finally by iron-rich coatings (Fig. 5s).
Chronological data
AMS radiocarbon ages
The calibrated 14C ages of mollusc shells ranged from ca. 30 ka to ca. 17 ka (Fig. 6a, Table S1.1), but exhibit substantial scatter and age inversions. Bioturbation and sediment reworking through slope wash and deflation likely contributed to this scatter. Overall, the chronological data indicate deposition of the upper unit starting from the end of MIS 3 and continuing during MIS 2, which is consistent with other regional data (Bosq et al., Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b), as well as data from elsewhere in Europe (Antoine et al., Reference Antoine, Coutard, Guerin, Deschodt, Goval, Locht and Paris2016; Lehmkuhl et al., Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021).
Figure 6. Chronological data of the Baix loess-paleosol section (LPS), Rhône Rift Valley, SE France. (A) Upper 7 m of main profile Baix with results of 14C ages of mollusc shells and sample positions of OSL7 and OSL8 taken from profile Baix_west (green squares). (B) Lower 7 m of main profile Baix with results of the OSL dating on feldspar coarse grains (pIR60IR225, green squares), polymineral fine grains (pIR60IR225, red circles), and quartz fine grains (BLSL, blue filled circles) from six OSL block samples (OSL 12020–OSL 32020 and OSL 62018–OSL 82018) of profiles Baix and Baix_west. The yellow squares mark the positions of the dose rate samples for the OSL screening. For comparison, OSL screening results presented in Pfaffner et al. (Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024) displayed with gray diamonds. Solid vertical lines denote the boundaries of the Marine Isotope Stages (MIS), and dotted lines mark the peaks of the MIS 5 substages (MIS boundaries after Lisiecki and Rayno, Reference Lisiecki and Raymo2005; Railsback et al., Reference Railsback, Gibbard, Head, Voarintsoa and Toucanne2015).
OSL ages
Both pIR60IR225 feldspar coarse-grain and pIR60IR225 polymineral fine-grain ages were well in line with the results of the OSL screening presented in Pfaffner et al. (Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024) (Fig. 6b). Therefore, the results of the pIR60IR225 dating provide methodically reliable single-point data that were complemented by the closely spaced OSL screening data providing a quasi-continuous chronometric framework. The course of the OSL screening data and the pIR60IR225 ages met the stratigraphic expectations derived from fieldwork and pedosedimentary analyses (cf., Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024).
In contrast, the quartz fine grain data produced age overestimates throughout the profile. This finding is surprising, because the luminescence signal of quartz is assumed to reset faster than that of feldspar (Godfrey-Smith et al., Reference Godfrey-Smith, Huntley and Chen1988). Additional investigations of the quartz fine grains showed that the thermally stimulated (TL) glow curves of the natural samples, but not of the artificially regenerated samples, exhibited a high-temperature peak, similar to a Precambrian quartz sample with a geologic dose suggesting a minimum age of ca 1.5 Ga (published by Schmidt and Woda, Reference Schmidt and Woda2019). This finding corroborated the assumption that geologically old fine-grained quartz from the underlying bedrock (Cretaceous marlstone) was incorporated into the Baix loess deposits, likely during slope wash processes.
Therefore, we disregarded the quartz fine grain ages based our interpretation of the results of the OSL screening (Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024), in combination with the pIR60IR225 dating of the feldspar coarse grains and the polymineral fine grains (Fig. 6b; Supplement 2: S2.6 additional tables). As, by nature, the coarse grain ages come with a smaller error than the polymineral fine-grain ages, in the discussion we refer to the pIR60IR225 feldspar coarse grain ages.
In the interpretation and contextualization of our results, we correlated the different phases of loess deposition and soil formation with Marine Isotope Stages (MIS) assuming stage boundaries according to Lisiecki and Raymo (Reference Lisiecki and Raymo2005) and Railsback et al. (Reference Railsback, Gibbard, Head, Voarintsoa and Toucanne2015). Therefore, MIS correspond to the terrestrial periods of the last interglacial (Eemian, MIS 5e), last early glacial (MIS 5a–d), early pleniglacial (MIS 4), middle pleniglacial (MIS 3), and late pleniglacial (MIS 2) periods (Figs. 6a, 6b, 7). Possible Greenland stadials (GS) and interstadials (GI) of the INTIMATE event stratigraphy (Rasmussen et al., Reference Rasmussen, Bigler, Blockley, Blunier, Buchardt, Clausen and Cvijanovic2014) are cautiously proposed, especially for periods of soil formation related to the paleosol horizons, as they are expected to respond to more supra-regional climate forces.
Discussion
Inference of sedimentation dynamics from granulometric characteristics
The peak in the coarse silt fraction point to one regional source, most probably the riverbed of the Rhône River (Fig. 4A). This agrees with Bosq et al. (Reference Bosq, Bertran, Degeai, Queffelec and Moine2020a, Reference Bosq, Bertran, Degeai, Queffelec and Moineb), who suggested that the riverbed of the Rhône River, carrying meltwater and abundant outwash sediments from the Rhône Glacier during the last glacial period, was the main source of loess in Rhône Rift Valley. The location of the loess deposit at the Baix LPS, which is situated on the foot slope of a small ridge in a leeward position with respect to the strong northern wind (Mistral) blowing down the Rhône Rift Valley, supports this assumption (Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024).
Bosq et al. (Reference Bosq, Bertran, Degeai, Kreutzer, Queffelec, Moine and Morin2018) generally attributed the change from silty loess north of Valence to sandy loess in the south to the topography of the Rhône Rift Valley. The channeled and accelerated wind transported sand- and silt-sized particles from the Rhône River south at the outlet of the valley, where the accumulation area is characterized by deposition of loess, cover sands and/or restricted/limited dune fields (Bosq et al., Reference Bosq, Bertran, Degeai, Kreutzer, Queffelec, Moine and Morin2018, Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b; Bertran et al., Reference Bertran, Bosq, Borderie, Coussot, Coutard, Deschodt, Franc, Gardère, Liard and Wuscher2021). Thus, the deposits of Baix LPS correspond to the northern (silty) loess, described by Lehmkuhl et al. (Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021) as subdomain IIIa (no sandy loess or aeolian sand).
Two subordinate maxima in the fine- to medium-silt and in the medium-clay fractions were identified (Fig. 4B). The observed tailing in the medium- to fine-silt and clay fractions may be attributed either to long-term suspension in the atmosphere, temporarily weakening wind activity, or to aeolian transport of silt-sized clayey aggregates (Tsoar and Pye, Reference Tsoar and Pye1987; Pye, Reference Pye1995; Vandenberghe, Reference Vandenberghe2013). It also may have been caused by post-depositional weathering (Bronger, Reference Bronger1976; Újvári et al., Reference Újvári, Kok, Varga and Kovács2016a; Schulte et al., Reference Schulte, Sprafke, Rodrigues and Fitzsimmons2018; Bosq et al., Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b; Sprafke et al., Reference Sprafke, Schulte, Meyer-Heintze, Händel, Einwögerer, Simon, Peticzka, Schäfer, Lehmkuhl and Terhorst2020), because pedogenesis in loess leads to an increase in medium to fine silt and clay at the expense of coarse silt (Sun et al., Reference Sun, Lu and An2000; Günster et al., Reference Günster, Eck, Skowronek and Zöller2001; Újvári et al., Reference Újvári, Kok, Varga and Kovács2016a; Wacha et al., Reference Wacha, Rolf, Hambach, Frechen, Galović and Duchoslav2018). The latter explanation seems the most likely in this case, because other observations, such as the very diffuse upper boundary of the EPG-PS, also indicate accretional soil formation. This points to an increase in aeolian input while weathering proceeded (Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024). This alteration effect was also reflected in the particle size distribution characteristics of the paleosol horizons and recent soil horizon of the Baix LPS, with a main mode in the coarse-silt fraction, high fine- and medium-silt contents in the Bw horizons, and increased fine clay contents in the Bt and Btg horizons (Fig. 4B).
Furthermore, the loess and reworked loess from bottom to top showed an increase in GSI and U-ratio, a decrease in the silt content, and an increase in fine sand content (Fig. 4A). Similar trends of upward coarsening were observed in the Lautagne LPS (Bosq et al., Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b), possibly the Val Sorda LPS (Ferraro, Reference Ferraro2009), in the LPS of Susak, Croatia (Wacha et al., Reference Wacha, Rolf, Hambach, Frechen, Galović and Duchoslav2018), and in central European LPSs (Antoine et al., Reference Antoine, Rousseau, Zöller, Lang, Munaut, Hatté and Fontugne2001, Reference Antoine, Rousseau, Moine, Kunesch, Hatté, Lang, Tissoux and Zöller2009; Schirmer, Reference Schirmer2016; Fischer et al., Reference Fischer, Hambach, Klasen, Schulte, Zeeden, Steininger, Lehmkuhl, Gerlach and Radtke2019). In other studies, such coarsening has been explained by increasing local wind intensity and/or by a shift from a more-distant to a somewhat nearer sediment source (Rousseau et al., Reference Rousseau, Antoine, Hatté, Lang, Zöller, Fontugne and Othman2002; Antoine et al., Reference Antoine, Rousseau, Moine, Kunesch, Hatté, Lang, Tissoux and Zöller2009; Wacha et al., Reference Wacha, Rolf, Hambach, Frechen, Galović and Duchoslav2018). In the case of the Baix LPS, increasing wind velocity towards the last glacial maximum (LGM) seems more likely than a change in the source material, which is assumed to be the riverbed of the Rhône River throughout the entire Late Pleistocene. Another possible explanation is the assumed persistence of a vegetation cover (Tsoar and Pye, Reference Tsoar and Pye1987; Bosq et al., Reference Bosq, Bertran, Degeai, Kreutzer, Queffelec, Moine and Morin2018; Bertran et al., Reference Bertran, Bosq, Borderie, Coussot, Coutard, Deschodt, Franc, Gardère, Liard and Wuscher2021). Due to its southern location, the Rhône Rift Valley is thought to have served as a refugium for trees and shrubs during the last glacial maximum, as it was still somewhat milder than its higher-lying surroundings (Beaudouin et al., Reference Beaudouin, Jouet, Suc, Berné and Escarguel2007; Tzedakis et al., Reference Tzedakis, Emerson and Hewitt2013). This vegetational influence was also implemented in the sediment depositional models suggested by Bosq et al. (Reference Bosq, Bertran, Degeai, Kreutzer, Queffelec, Moine and Morin2018) (Fig. 11) and Bertran et al. (Reference Bertran, Bosq, Borderie, Coussot, Coutard, Deschodt, Franc, Gardère, Liard and Wuscher2021), who assumed a prominent link between the steep relief and taller vegetation cover trapping loess particles more effectively in the Rhône Rift Valley.
Geochemistry of Baix LPS and paleoenvironmental interpretation
Characteristics of the non-weathered or only slightly weathered units of loess and reworked loess
Based on field observations (Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024), the PSD, and geochemical characteristics, the C horizons (characterized by maxima of U-ratio, GSI, and primary carbonate content and minima of b values, weathering indices, and Mnd/Mnt ratios) represent the “least weathered sediment units” of the Baix LPS (Figs. 3, 4). This generally points to increased sediment accumulation rates, reflecting the coldest and driest conditions. In contrast, the CB and BC horizons (with slightly decreased CaCO3 content and WIMER values, and increased CPA values) represent shifts towards somewhat milder and probably more humid climate conditions.
Both categories (5 BCk5 to 5 Ck2 and 3 BCk3 to 2 BCk2 to CBk to Ck1) were marked by bioturbation and carbonate leaching indicated by secondary carbonate nodules that mostly showed no distinct depth of accumulation, as well as gradual changes in the weathering ratios and color parameters a and a/b. Collectively, this indicates increasing predominance of loess accumulation over weathering, and possibly decreasing humidity that slowed decarbonization and weathering processes.
LPSs from northern regions indicate conditions related to open vegetation with almost no trees (Gocke at al., Reference Gocke, Hambach, Eckmeier, Schwark, Zöller, Fuchs, Löscher and Wiesenberg2014), the increased SOC content within the C horizons together with evidence of synsedimentary bioturbation of Baix LPS may be interpreted as reflecting persisting vegetation at the site and organic matter accumulation even during the coldest phases (Fig. 3). However, the SOC contents reported here should not be over interpreted, since they come with a considerable analytical error as explained in the methods section.
In contrast to such gradual changes, there were also two horizons (6 BCk/Btg and 4 BCk/Bw), in which material of an underlying paleosol horizon was mixed with fresh loess. These horizons were characterized by abrupt increases in the U-ratio, GSI, and Si/Al, as well as lower a and a/b ratio, lower CPA and Fed/Fet ratio, and a higher WIMER (Fig. 3). All of these indicate abrupt climatic deteriorations at these two points in time, which triggered erosion and re-deposition of soil material, while aeolian processes intensified. These deteriorations (colder, drier conditions) are characterized by decreased evapotranspiration, also shown by a decline of temperate forest vegetation towards tundra/steppe vegetation (Reille and de Beaulieu, Reference Reille and de Beaulieu1990; Beaudouin et al., Reference Beaudouin, Suc, Acherki, Courtois, Rabineau, Aloïsi, Sierro and Oberlin2005; Mologni et al., Reference Mologni, Purdue, Audiard, Dubar, Kreutzer and Texier2021).
Throughout the entire profile, no lamella, platy structure, cracks, ice-wedge pseudomorphs, or any other features related to severe frost-thaw were observed. This observation is in accordance with the suggested maximum southernmost occurrence of permafrost at 47°N during the coldest period of the last glacial period (does not strictly coincide with the LGM) (Andrieux et al., Reference Andrieux, Bertran and Saito2016). Thus, frost-induced mass movement (e.g., solifluction) processes in Baix could only have taken place during early spring, above the temporarily frozen subsoil. The main sediment reworking along the slope most likely took place through slope wash processes (e.g., during snowmelt). In the relief position of the Baix LPS, on the foot slope of a small ridge, only short-distance slope wash processes were possible (Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024; Fig. 2).
Characteristics of the paleosols
Basal last interglacial to last early glacial paleosol complex (IG-EG-PSC)
Generally, the horizons of the basal IG-EG-PSC were characterized by complete decarbonization and an advanced stage of chemical weathering and pedogenesis, as indicated by the highest Fed/Fet ratios and CPA values, and lowest WIMER values and Si/Al ratios. In addition, it showed a reddish color, as reflected in highest a values and a/b ratios, interpreted as slight rubefaction (Fig. 3). The lowermost horizon, 8 Bt2, developed in a slope deposit of weathered material of the underlying bedrock, which explains its comparatively high clay content (Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024). The overlying 7 Bt1 and 7 Btg horizons showed tonguing, generally upward increasing CPA values, Fed/Fet, and Mnd/Mnt ratios and a values, and decreasing WIMER values. Highest fine-clay/coarse-clay ratios and clay coatings indicate clay illuviation (Fig. 4A, 4B). All these characteristics pointed to warm and humid conditions during the formation of this paleosol complex. The slight rubefaction of the paleosol complex matches with that of the last interglacial paleosol of central European LPSs. For instance, Antoine et al. (Reference Antoine, Rousseau, Zöller, Lang, Munaut, Hatté and Fontugne2001) reported a “brown red loam” in the last interglacial paleosol at Nussloch (unit 2b). In contrast, strong rubefaction generally has been reported from last interglacial paleosols in Mediterranean LPSs (Ferraro, Reference Ferraro2009; Wacha et al., Reference Wacha, Mikulčić Pavlaković, Novothny, Crnjaković and Frechen2011b; Boixadera et al., Reference Boixadera, Poch, Lowick and Balasch2015, Durn et al., Reference Durn, Rubinić, Wacha, Patekar, Frechen, Tsukamoto, Tadej and Husnjak2018; Zhang et al., Reference Zhang, Rolf, Wacha, Tsukamoto, Durn and Frechen2018; Bosq et al., Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b). There are two possible explanations for the only slightly rubified last interglacial paleosol complex in the Baix LPS: (1) the last interglacial climatic conditions at Baix were similar to those in central Europe; and (2) the position of the Baix LPS at the foot slope of a small ridge resulted in increased moisture that prevented the soil from drying out during summer and, thus, hampered the rubefaction process. So far, no other last interglacial paleosol has been studied at this climatic transition. However, Mediterranean climate conditions during the last interglacial have been described from other nearby archives (de Beaulieu and Reille, Reference de Beaulieu and Reille1984; Tzedakis, Reference Tzedakis2007; Valladas et al., Reference Valladas, Mercier, Ayliffe, Falguères, Bahain, Dolo and Froget2008; Rivals et al., Reference Rivals, Moncel and Patou-Mathis2009; Moncel et al., Reference Moncel, Allué, Bailon, Barshay-Szmidt, Béarez, Crégut and Daujeard2015). Thus, given the foot slope position of the Baix LPS and high clay content of its lowermost horizon (8 Bt2), favoring interflow along the slope, the inhibition of rubefaction by enhanced soil moisture in this foot slope position seems the more likely explanation.
The comparatively weakly developed soil structure of the 7 Btg horizon, and the co-occurrence of in-situ and reworked clay coatings point to the polygenetic nature of the paleosol complex. In a first phase of pedogenesis, a Luvisol formed, from which only the 7 Bt1 and 8 Bt2 horizons have been preserved. The Luvisol E horizon was completely eroded, and the upper part of the Bt horizon was reworked, mainly by slope processes. Thus, climatic deterioration must have opened the vegetation cover to some extent. The reworked material again was subjected to clay migration, while enhanced moisture developed hydromorphic features, resulting in the formation of the 7 Btg horizon. This reworking and subsequent clay illuviation are indicated by the co-occurrence of broken and intact clay coatings (Fig. 5d, q, s). In addition, the formation of the stagnic color pattern might have been enhanced as the repeated clay illuviation reduced the hydraulic conductivity over time. The strong hydromorphic features, also reported from LPS between Cadarache and Saint-Estève (Durance Valley, southern Rhône Rift Valley) (Bonifay, Reference Bonifay1965; Mologni et al., Reference Mologni, Purdue, Audiard, Dubar, Kreutzer and Texier2021), most likely reflect climatic shift towards more humid and probably cooler conditions. The observed clay–silt coatings in the absence of permafrost indicators point to the regular melting of thick annual snow covers (Fedoroff and Goldberg, Reference Fedoroff and Goldberg1982). This scenario seems likely for the phase when sclerophyllous taxa declined at the end of the last interglacial towards temperate deciduous forest during the following interstadial periods (MIS5 c and a) (de Beaulieu and Reille, Reference de Beaulieu and Reille1984; Reille and de Beaulieu, Reference Reille and de Beaulieu1990; Mologni et al., Reference Mologni, Purdue, Audiard, Dubar, Kreutzer and Texier2021).
Last early pleniglacial paleosol (EPG-PS)
The EPG-PS was represented by a decarbonated bipartite brown 6 Bw3 horizon (Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024). Its weathering ratios and increased fine silt proportion (∼5–12 µm) reflect a moderate stage of pedogenesis (Figs. 3, 4). The lower part of the 6 Bw3 horizon still comprised some relocated reddish aggregates and depletion zones that had apparently been mixed in from the 7 Btg horizon of the IG-EG-PSC, as well as in-situ clay–silt coatings (Fig. 5c, k). Thus, the Bw formation must have started by formation of the 6 Bw3 horizon, which is a mixture of fresh sediment and pre-weathered soil material relocated from the underlying 7 Btg horizon. Pedogenesis still continued, when slow loess accumulation set in, leading to a gradual upward growth of the EPG-PS and formation of the 5 BCk5 horizon. In other words, the A horizons seemed to be transformed into B horizons as parts of the former A horizons were marked by decreasing SOC contents and an increase of biological activity. This process of accretional soil formation was also described by Kemp et al. (Reference Kemp2001) and Constantin et al. (Reference Constantin, Mason, Veres, Hambach, Panaiotu, Zeeden and Zhou2021) for the Pleistocene–Holocene climatic transition. Apparently, climatic conditions during this period were still humid enough to allow the formation of clay–silt coatings.
Middle pleniglacial paleosol complex (MPG-PSC)
The MPG-PSC developed in a reworked loess sediment, as reflected by high variability of the clay, silt, and sand contents (Fig. 4B). The decarbonated 5 Bw2 horizon displayed similar WIMER, CPA, clay contents, fine-silt proportion, U-ratio, GIS, and silt–clay coatings along pores as the 6 Bw3 horizon of the EPG-PS (Figs. 3, 4A, 5b, j). However, in contrast to the EPG-PS, we assume a top-down soil formation, where a substantial part of the MPG-PSC was eroded. The erosion is indicated by the large carbonate nodules in the 5 BCk4 horizon, which can only be explained by a much thicker decarbonated soil horizon, which must have been overlying the preserved remains of the 5 Bw2 horizon that shows a major erosional discontinuity on top.
The MPG-PSC in the Baix LPS was very similar to the one in the Collias LPS farther south (Bosq et al., Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b). In addition to the large carbonate nodules that are mainly restricted to one depth (the 5 BCk5 horizon), this paleosol in both LPSs showed smaller carbonate nodules that were distributed over a rather large depth range. At Collias, Bosq et al. (Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b) explained this observation by an interplay between relatively intensive soil formation and moderate loess deposition rates, similar to the concept of accretional soil formation (Kemp, Reference Kemp2001; Constantin et al., Reference Constantin, Mason, Veres, Hambach, Panaiotu, Zeeden and Zhou2021). With progressive loess input, both the soil surface and the corresponding carbonate accumulation zone moved upwards. In this way, the secondary carbonate nodules were distributed over a wide depth range (Bosq et al., Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b).
However, it must be considered that the MPG-PSC reflects successive events of pedogenesis and soil erosion, yet the complete decarbonization of the 5 Bw2 horizon and the occurrence of the large carbonate nodules at specific depths, both at Collias and at Baix, indicated at least one phase of increased humidity with soil formation without relevant loess input. However, the most strongly developed upper part of the MPG-PSC, from where the carbonates that now form the large nodules in the 5 BCk5 horizon were leached, has been eroded in both LPSs, indicating cross-regional climatic deterioration within the middle pleniglacial period.
However, it is challenging to identify a separate upper MIS 3 paleosol above the MPG-PSC (3 BCk3, 2 BCk2 horizons), because the brownish color of the slope wash loess deposits in that stratigraphic position could indicate either accretional soil formation (less intense than EPG-PS) or derivation from brown material eroded from the MPG-PSC in an upslope position and mixed in at the foot slope position of the Baix LPS. This interpretation is supported by the considerable variability in silt and sand contents within the 3 BCk3 horizon.
Holocene soil
The recent soil at Baix is an Endocalcic Cambisol. Compared to the paleosol Bw horizons (6 Bw3, 5 Bw2), the Bw1 horizon of the recent soil is not completely decarbonated, and its weathering ratios reflect a lower degree of weathering (Fig. 5a, g). This difference most likely is not a result of less moisture and lower weathering intensity during the Holocene period, but of rejuvenation of the soil through erosion. Although the site is forested at present, human influence is obvious (with an old quarry and agricultural activities nearby).
Chronology of Baix LPS
Last interglacial and last early glacial period (MIS 5)
The loess unit, in which the IG-EG-PSC formed (OSL 12020) gave a pIRIR-age of 117.5 ± 9.7 ka (Fig. 6b). However, because of the hydromorphic features in this paleosol, Pfaffner et al. (Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024) assumed a higher soil water content of Δ 1.28 ± 0.05 for the lowermost part of the main profile. Adopting this value for the present study yielded an age of 128.5 ± 10.5 ka, supporting the field assumption of pre-last interglacial loess deposition and soil formation during the last interglacial (ca. 126–115 ka; in the Mediterranean likely until 110 ka, according to Sánchez Goñi et al., Reference Sánchez Goñi, Eynaud, Turon and Shackleton1999) soil formation. Thus, formation of the 7 Bt1 horizon correlates with MIS 5e (GI 26/25c; Eemian).
The wide scatter of OSL screening data obtained for sediment of the overlying 7 Btg horizon is due to the reworking of MIS 5e soil material, involving partial bleaching of feldspar grains (Fig. 6b). The reworking phase and subsequent soil formation phase most likely took place during MIS 5d or MIS 5b and MIS 5c or MIS 5a, respectively (Fig. 7).
Figure 7. Compiled marine and ice-core chronostratigraphy for the last 140 ka, including left: Marine Isotope Stages (MIS) derived from global benthic stack (Lisiecki and Raymo, Reference Lisiecki and Raymo2005) and 1000-year average values of δ18O from NGRIP records including Greenland Interstadials (GI) from INTIMATE event stratigraphy (Rasmussen et al., Reference Rasmussen, Bigler, Blockley, Blunier, Buchardt, Clausen and Cvijanovic2014) and Heinrich events (H) (Allard et al., Reference Allard, Hughes and Woodward2021). Right: Generalized loess paleosol-sequence (LPS) of Nussloch (Antoine et al., Reference Antoine, Rousseau, Zöller, Lang, Munaut, Hatté and Fontugne2001, Reference Antoine, Rousseau, Moine, Kunesch, Hatté, Lang, Tissoux and Zöller2009; Kadereit et al., Reference Kadereit, Kind and Wagner2013; Moine et al., Reference Moine, Antoine, Hatté, Landais, Mathieu, Prud’homme and Rousseau2017), Baix and Collias LPS (Bosq et al., Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b, Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024) emphasizing soil-formation phases during chronostratigraphical stages of the Late Pleistocene with their latitudinal position, recent mean annual temperature (MAT), mean annual precipitation (MAP). and number of arid months (dwd, deutscher Wetterdienst, Reference dwd2021a, Reference dwdb; Météo-France 2025a, c).
Early pleniglacial period (MIS 4)
The OSL block sample (OSL 22020) taken from the 6 BCk/Btg horizon, directly overlying the prominent erosional discontinuity above the MIS 5 paleosol complex, gave an age of 67.0 ± 5.7 ka (Fig. 6b). Thus, that erosional discontinuity and overlying reworked sediment (6 BCk/Btg horizon) apparently mark climatic deterioration at the onset of the MIS 4 (71–57 ka).
Early to middle pleniglacial period (MIS 4 to MIS 3)
The next-upper block sample (OSL 32020) was taken from the only slightly weathered unit (5 BCk5 horizon) above the MIS 4 Calcic Cambisol. It yielded an age of 59.8 ± 5.3 ka, roughly corresponding to the MIS 4/MIS 3 boundary around 57 ka (Fig. 6b). Thus, the accretional soil formation of the Calcic Cambisol (6 Bw3) may have continued during the milder excursions of GI 18 until GI 16/15 (Fig. 7) and must have ended at the latest by the end of MIS 4. Deposition of the overlying sediment unit probably took place mainly during the cool periods between GS 18 and GS 15 (Fig. 7).
The sedimentary material in which the truncated prominent MIS 3 Calcic Cambisol had formed, showed a pIRIR age of 54.8 ± 5.2 ka (OSL 62018). This age overlaps within error margins with the age of 59.8 ± 5.3 ka obtained for the OSL 32020, taken from the sediment unit below, which may suggest a very high sedimentation rate. Considering the minimum of the error margins, sedimentation rates between the times yielded from samples OSL 32020 and OSL 62018 were possibly a factor 3.6 higher compared to those between the times given by samples OSL 22020 and OSL 32020. In contrast, considering the maximum of the error margins, it is also possible that the sediment units bracketed by samples OSL 32020 and OSL 62018 (in which the 5 BCk5 and 5 Bk2 developed) underwent a relatively short period of intensive sediment reworking that almost homogenized the pIRIR age of the whole unit. Because of the homogeneity of the material, field observation can neither confirm nor exclude this possibility. This observation of a possibly quasi-constant luminescence age of about 55 ka over a certain sediment depth is very similar to the situation at the Collias LPS, where OSL ages of 55.5 ± 4.4 ka and 55.1 ± 4.1 ka bracketed a ca. 2-m-thick sediment unit, in whose upper part the prominent MIS 3 Calcic Cambisol had formed (Bosq et al., Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b). Thus, the phenomenon may have occurred at the same time in the two LPSs.
Based on the sedimentation times of around 55 ka obtained in both LPSs, the formation of the MIS 3 Calcic Cambisol with the prominent carbonate nodules most likely started in the relatively long-lasting GI 14 (Fig. 7). The end of its pedogenesis and its truncation by severe erosion must pre-date the age of the overlying slope washed, weathered loess deposits (3 BCk3), which yielded an age of 37.5 ± 3.1 ka in the Baix LPS (OSL 72018) and an age of 38.5 ± 2.5 ka in the Collias LPS. Thus, the climatic deterioration that caused the end of the period of geomorphological stability during which the Calcic Cambisol formed, may correspond to GS 9 and/or the Heinrich 4 event (40.2–38.3 ka) (Fig. 7). Hence, the truncated MIS 3 Calcic Cambisol with the prominent carbonate nodules that was present in both LPSs, Baix and Collias, represents a pedo-complex that probably developed in repeated pedogenic cycles over a period that roughly includes GI 14–GI 9, representing milder phases during that time, which allowed for enhanced soil formation.
Middle to late pleniglacial period (MIS 3 to MIS 2)
Another time mark of ca 35.8 ± 3.2 ka was obtained from OSL 82018. Together with the ages obtained from the next-lower OSL 72018 (37.5 ± 3.1 ka) and OSL 62018 (54.8 ± 5.2 ka), this age supported a trend of upward-decreasing sedimentation ages for the uppermost ∼4.5 m of the LPS (Fig. 6a, b).
This was important information, because the ages from the OSL screening showed such scatter for this part of the LPS that it was difficult to obtain an age trend from them (Pfaffner et al., Reference Pfaffner, Kadereit, Karius, Kolb, Kreutzer and Sauer2024). AMS 14C ages of mollusc shells from the same part of the LPS were much younger, corresponding to late MIS 3 and MIS 2 (Fig. 6a). They showed a trend of decreasing ages but with some age inversions.
Comparison of Baix LPS with other European LPSs
Polygenetic Luvisol formation during the last interglacial (MIS 5e) to last early glacial (MIS 5d-5a) periods also has been reported from various central European and Mediterranean LPSs (Bourdier, Reference Bourdier1958; Bonifay, Reference Bonifay1965; Antoine et al., Reference Antoine, Rousseau, J-P and Hatté1999, Reference Antoine, Rousseau, Zöller, Lang, Munaut, Hatté and Fontugne2001; Rose et al., Reference Rose, Meng and Watson1999; Günster et al., Reference Günster, Eck, Skowronek and Zöller2001; Ferraro, Reference Ferraro2009; Wacha et al., Reference Wacha, Mikulčić Pavlaković, Frechen and Crnjaković2011a, Reference Wacha, Mikulčić Pavlaković, Frechen and Crnjakovićb; Boixadera et al., Reference Boixadera, Poch, Lowick and Balasch2015; Haesaerts et al., Reference Haesaerts, Damblon, Gerasimenko, Spagna and Pirson2016; Bosq et al., Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b; Kehl et al., Reference Kehl, Seeger, Pötter, Schulte, Klasen, Zickel, Pastoors and Claßen2024). At Baix LPS, we assume a warm interglacial climate during MIS 5e and at least one period of sediment reworking, followed by a period of soil development involving the formation of hydromorphic features during MIS 5d–a. Similar pedo-stratigraphic units for the MIS 5 have been described from LPSs of southern France (Bosq et al., Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b), Spain (Lachar LPS; Günster et al., Reference Günster, Eck, Skowronek and Zöller2001), and Croatia (Susak LPSs; Wacha et al., Reference Wacha, Mikulčić Pavlaković, Novothny, Crnjaković and Frechen2011b; Profe et al., Reference Profe, Wacha, Frechen, Ohlendorf and Zolitschka2018). In contrast, central European LPSs reflect more continental (drier) conditions during the MIS 5 interstadial periods that are typically represented by up to three humus horizons (e.g., the Mosbach humus zones, the Bettencourt soil overlying the St. Sauflieu complex, the upper Rocourt pedocomplex overlying the humus complex of Remicourt, and the upper part of the Garzweiler pedocomplex) (Antoine et al., Reference Antoine, Rousseau, J-P and Hatté1999, Reference Andrieux, Bertran and Saito2016; Schirmer, Reference Schirmer, Ikinger and Schirmer2002, Reference Schirmer2016; Frechen et al., Reference Frechen, Terhorst and Rähle2007; Lehmkuhl et al., Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021).
The erosional discontinuity above the IG-EG-PSC of Baix LPS apparently marks the climatic deterioration at the onset of the MIS 4 (71–57 ka). In the western and central European loess belt, this period is known for widespread sediment reworking (e.g., Keldach discordance and Niedereschbach zone; Hermies, Harmignies colluvial deposit) (Semmel, Reference Semmel and Ikinger1998; Antoine et al., Reference Antoine, Coutard, Guerin, Deschodt, Goval, Locht and Paris2016; Lehmkuhl et al., Reference Lehmkuhl, Zens, Krauß, Schulte and Kels2016). However, for the Rhône Rift Valley, this is the first time that this landscape instability at the onset of MIS 4 has been precisely identified and dated.
So far, MIS 4 paleosols such as the EPG-PS of Baix have been reported from very few Mediterranean LPSs. Thin brown MIS 4 paleosols were observed in the Bok section in Croatia (Wacha et al., Reference Wacha, Mikulčić Pavlaković, Frechen and Crnjaković2011a). An orange-brown paleosol of the Bok section 1 (Wacha et al., Reference Wacha, Mikulčić Pavlaković, Novothny, Crnjaković and Frechen2011b) that developed after 98.3 ± 7.3 ka and prior to 54.3 ± 4.9 ka might correspond to the Calcic Cambisol in the Baix LPS. In central Europe, MIS 4 to MIS 4/MIS 3 Cambisol and Regosol relics (e.g., the Gräselberg Soil (Gräselberg LPS; Semmel, Reference Semmel1968); the Jackerath Soil and Reisberg Soil (Garzweiler and Schwalbenberg I LPS; Schirmer, Reference Schirmer, Ikinger and Schirmer2002), the Havrincourt Soil (Havrincourt LPS; Antoine et al., Reference Antoine, Coutard, Guerin, Deschodt, Goval, Locht and Paris2016), the Malplaquet Soil (Remicourt, Harmignies LPSs; Haesaerts et al., Reference Haesaerts, Damblon, Gerasimenko, Spagna and Pirson2016), stratigraphic units 12, 20, 23, and possibly 26 – all Bw horizons of Calcaric Cambisols (Schwalbenberg LPS; Fischer et al., Reference Fischer, Jöris, Fitzsimmons, Vinnepand, Prud’homme, Schulte and Hatté2021) have been reported from several LPSs. The temporal positioning of some of these paleosols seems clear (e.g., that of the Jackerath Soil in the MIS 4), while the allocation of others appears less unequivocal, but is probably somewhere near the MIS 4/MIS 3 boundary (e.g., in the case of the Lower Gräselberg Soil; Antoine at al., Reference Antoine, Rousseau, Moine, Kunesch, Hatté, Lang, Tissoux and Zöller2009; the Lower Remagen Soils; Schirmer, Reference Schirmer2012), possibly corresponding to the brown Calcic Cambisol in the Baix LPS (Fig. 7).
An intensively red-brown soil horizon, underlain by carbonate nodules, similar to the truncated MPG-PSC of Baix LPS has been described on Susak Island (Wacha et al., Reference Wacha, Mikulčić Pavlaković, Novothny, Crnjaković and Frechen2011b, Reference Wacha, Rolf, Hambach, Frechen, Galović and Duchoslav2018; Zhang et al., Reference Zhang, Rolf, Wacha, Tsukamoto, Durn and Frechen2018). Furthermore, a comparably well-developed brown, truncated paleosol horizon (of a Cambisol) is found in most European LPSs (although without carbonate nodules) in northern and southern France (Saint-Acheul/Villiers-Adam soil complex) (Bonifay, Reference Bonifay1965; Antoine et al., Reference Antoine, Bahain, Debenham, Frechen, Gauthier, Hatté, Limondin-Lozouet, Locht, Raymond and Rousseau2003, Reference Andrieux, Bertran and Saito2016), in Belgium (Les Vaux) (Haesaerts et al., Reference Haesaerts, Damblon, Gerasimenko, Spagna and Pirson2016), and in Germany, where sometimes even several brown paleosol horizons have been preserved in this stratigraphic position (e.g., Remagen Soils, Gräselberg Soils) (Antoine et al., Reference Antoine, Rousseau, Zöller, Lang, Munaut, Hatté and Fontugne2001; Schirmer, Reference Schirmer, Ikinger and Schirmer2002, Reference Schirmer2016; Frechen et al., Reference Frechen, Terhorst and Rähle2007; Gocke et al., Reference Gocke, Hambach, Eckmeier, Schwark, Zöller, Fuchs, Löscher and Wiesenberg2014; Fischer et al., Reference Fischer, Jöris, Fitzsimmons, Vinnepand, Prud’homme, Schulte and Hatté2021). Thus, the truncated Calcic Cambisol with the prominent carbonate nodules at Baix may be regarded as a mid-MIS 3 stratigraphic marker within the Rhône Rift Valley that also may serve for cross-regional correlation of central and south European LPSs.
Significant soil formation was not detected above the truncated MPG-PSC either at Baix LPS or at Collias LPSs. In Mediterranean LPSs, the late MIS 3 paleosols are highly diverse, depending on their geographical and geomorphological positions (Brunnacker, Reference Brunnacker1974; Coudé-Gaussen, Reference Coudé-Gaussen1990). They vary from up to three Chernozems in Italy (Ferraro, Reference Ferraro2009) to up to four thin, brown paleosol horizons in the Susak LPS (Wacha et al., Reference Wacha, Mikulčić Pavlaković, Frechen and Crnjaković2011a), one of the few LPSs with reliable age control for this period. For the northern Rhône Rift Valley, Franc et al. (Reference Franc, Moine, Fülling, Auguste, Pasty, Gadiolet, Gaertner and Robert2017) also described a possibly correlating brown paleosol, whereas in several LPSs of the southern Rhône Rift Valley the existence of a paleosol in this stratigraphic position is unclear (Bourdier, Reference Bourdier1958; Bonifay, Reference Bonifay1965), while in others such paleosols are missing (Bosq et al., Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b). In addition to the MIS 3 paleosols described above, the central European LPSs typically show one or several considerably younger brown paleosol horizon(s) that formed during the milder periods GI 8 to GI 4 (ca. 38–29 ka) making up the Lohne Soil or its equivalents (e.g., the Sinzig Soils, upper part of the Saint-Acheul/Villiers-Adam soil complex) (Antoine et al., Reference Antoine, Rousseau, Zöller, Lang, Munaut, Hatté and Fontugne2001; Schirmer, Reference Schirmer, Ikinger and Schirmer2002; Kadereit et al., Reference Kadereit, Kind and Wagner2013; Moine et al., Reference Moine, Antoine, Hatté, Landais, Mathieu, Prud’homme and Rousseau2017; Fischer et al., Reference Fischer, Hambach, Klasen, Schulte, Zeeden, Steininger, Lehmkuhl, Gerlach and Radtke2019).
Either possible younger MIS 3 incipient soils were later mixed in with younger loess material, or the Baix LPS does not conform with the central European LPS after ca. 40 ka, but rather conforms to the south European LPS, where climatic deterioration may have been more gradual compared to the more extreme climatic fluctuations of mid-European LPSs. However, the Baix LPS apparently experienced intense loess sedimentation for the period around 40–35 ka. A comparably high sedimentation rate occurred in the Collias LPS ca. 39–37 ka.
Such high loess accumulation rates that occurred at both sites might have been related to dry conditions during Greenland stadials and especially to Heinrich events, which caused a dramatic decline of forest and led to more open vegetation (Fletcher and Sánchez Goñi, Reference Fletcher and Sánchez Goñi2008). Dry conditions during Heinrich events are also assumed to have caused glacier retreat in the Mediterranean (Allard et al., Reference Allard, Hughes and Woodward2021) as well as high sediment load of Mediterranean rivers (e.g., Woodward et al., Reference Woodward, Hamlin, Macklin, Hughes and Lewin2008; Lewis et al., Reference Lewis, McDonald, Sancho, Peña and Rhodes2009), whose riverbeds provided major dust source areas when drying out. The study of Bosq et al. (Reference Bosq, Kreutzer, Bertran, Lanos, Dufresne and Schmidt2023) suggests an earlier start (approx. 42 ka – Alpine ice sheet) of enhanced loess deposition within the Alpine region compared to the mid-European region (approx. 32 ka – Fennoscandian ice sheet; Northern European Loess Belt). Thus, the lack of younger MIS 3 paleosols also may be explained by asynchronous loess deposits within the European aeolian system, in contrast to the previous assumed less-extreme climatic fluctuations for Baix LPS. The continuing loess accumulation observed at both LPSs, Baix and Collias, during MIS 2, matches with the acceleration of loess accumulation reported from many central European LPSs (Antoine et al., Reference Antoine, Rousseau, Moine, Kunesch, Hatté, Lang, Tissoux and Zöller2009, Reference Andrieux, Bertran and Saito2016; Ferraro, Reference Ferraro2009).
Conclusion
In this work, we characterized the loess and reworked loess, and the paleosol horizons of the Baix LPS based on granulometric, geochemical, and micromorphological analyses. We also established a detailed stratigraphy of the Baix LPS with a robust chronostratigraphic framework, including the last interglacial and glacial period (MIS 5 to MIS 2). The particle size distribution of the loess and reworked loess at Baix corresponds well with characteristics of central European loess. Loess accumulation at Baix was mainly controlled by the local wind regime, topography, and vegetation cover. The main loess accumulation phases occurred during the last early pleniglacial (MIS 4) and late pleniglacial (MIS 2) periods, whereas an earlier beginning at the end of the MIS 3 at Baix LPS compared to mid-European LPSs seems highly likely.
At the base of the Baix LPS, a well-preserved record of the last interglacial to last early glacial period (Eemian, MIS 5e) is provided by a polygenetic, slightly rubified stagnic Luvisol. It indicates warm last interglacial conditions, followed by a phase of erosion and reworking caused by climatic deterioration and a subsequent phase of pedogenesis under cooler and moister conditions (due to decreased evapotranspiration and annual snowmelt). The MIS 5 paleosol complex in the Baix LPS is much less rubified than its counterpart in the Collias LPS, although the Mediterranean climate most likely expanded as far north as Baix during MIS 5e. This is probably due to the strong hydromorphic processes that overprinted the paleosol complex in the Baix LPS and prevented more intense rubefaction (Yaalon, Reference Yaalon1997).
During the early pleniglacial period (MIS 4) accretional soil formation took place forming a Calcic Cambisol. Its brown Bw horizon is overlain by a BCk and a Ck (unweathered loess) horizon, reflecting increasing aeolian input while synsedimentary pedogenesis continued for a while and then weakened.
The middle pleniglacial period (MIS 3) allowed the formation of at least one well-developed Calcic Cambisol (top-down soil formation) with large carbonate nodules. This paleosol needed relatively mild and humid conditions to develop, which makes its development during GI 14 to GI 9 most likely. The Calcic Cambisol was then truncated, possibly during the climatic deterioration of GS 9.
During the late middle pleniglacial (MIS 3; from GS 9 on) and late pleniglacial (MIS 2) periods, loess accumulation prevailed. No soil horizons like those reported for MIS 2 in the central European LPSs (tundra gley soil; Antoine et al., Reference Antoine, Rousseau, Moine, Kunesch, Hatté, Lang, Tissoux and Zöller2009; cryosols and regosols; Lehmkuhl et al., Reference Lehmkuhl, Zens, Krauß, Schulte and Kels2016) were found at Baix LPS during this period. Considering an earlier beginning of pronounced loess accumulation, a milder climate at Baix, compared to central Europe, may also be a reason for the more pronounced bioturbation in the Baix LPS than in the central European LPSs. In this respect, the Baix LPS resembles the Mediterranean LPSs, for which gradual climatic changes are typical.
The recent Cambisol is only moderately developed, because anthropogenic disturbances caused its erosion and, hence, rejuvenation. Therefore, it cannot be compared to the well-developed MIS 5e paleosol, to evaluate the intensity of weathering and soil formation during MIS 5e conditions.
Overall, the Baix LPS displays characteristics of both the central European and south European LPSs and, thus, reflects its location in a climatic transition zone, as suggested by Brunnacker (Reference Brunnacker1974), Bosq et al. (Reference Bosq, Bertran, Degeai, Kreutzer, Queffelec, Moine and Morin2018, Reference Bosq, Kreutzer, Bertran, Degeai, Dugas, Kadereit and Lanos2020b), and Lehmkuhl et al. (Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021) (subdomain IIIa to VIa). This combination of characteristics makes the Baix LPS a valuable link between the loess regions of central and southern Europe.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/qua.2025.10053.
Acknowledgments
We are very grateful to Monsieur L. Garnier for allowing us to prepare the profile in his quarry. We thank Dr. Barbara von der Lühe for help in the field, and Dr. Volker Karius, Dr. Jürgen Grotheer, Anja Södje, Petra Voigt, Marius Friedrich, Jutta Asmuth, Dr. Simon Meyer-Heintze, and Dr. Phillip Schulte for the laboratory analyses and graphical support. We acknowledge the German Aerospace Centre (DLR), which provided the TanDEM-X data for the study area (DEM_GEOL1263). This work was funded by the Department of Physical Geography of the University of Goettingen. Sebastian Kreutzer was supported by the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement no. 844457 (CREDit).
Author contributions
Conceptualization of the project: Daniela Sauer and Annette Kadereit; field investigation: Nora Pfaffner, Daniela Sauer, Annette Kadereit, Tianhao Wang, Pascal Bertran, Mathieu Bosq; methodology: Nora Pfaffner, Annette Kadereit, Sebastian Kreutzer, Thomas Kolb, Hans-Peter Meyer, Alexander Varychev Pascal Bertran, Christine Hatté; data validation and interpretation: Nora Pfaffner, Daniela Sauer, Annette Kadereit, Sebastian Kreutzer Thomas Kolb, Pascal Bertran, Mathieu Bosq, Christine Hatté; conceptualization of the manuscript: Nora Pfaffner; writing – original draft preparation: Nora Pfaffner; writing – review and editing: all authors. All authors have read and agreed to the published version of the manuscript.
Conflicts of interest
The authors declare no conflicts of interest.
Open access
