Introduction
Loess is a sediment of aeolian genesis, deposited mainly during cold periods of the Quaternary; its composition is dominated by silt-sized quartz particles (Pye, Reference Pye1984, Reference Pye1995; Smalley et al., Reference Smalley, Jefferson, Dijkstra and Derbyshire2001). In Europe, loess occurs along and outside of the southern boundary of the Pleistocene Fennoscandian ice sheet, forming the Northern European Loess Belt (NELB), which stretches from southern Great Britain to northern France, Belgium, Netherlands, through Germany and Poland to the Eastern European Plain (Różycki, Reference Różycki1991; Pécsi and Richter, Reference Pécsi and Richter1996; Smalley and Jary, Reference Smalley and Jary2005; Haase et al., Reference Haase, Fink, Haase, Ruske, Pécsi, Richter, Altermann and Jäger2007; Bertran et al., Reference Bertran, Liard, Sitzia and Tissoux2016; Lehmkuhl et al., Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021). Climatic records derived from these deposits show a somewhat consistent pattern, with repeating marker horizons that enable connection of different sequences over long distances. Varying climatic and topographic conditions led directly to the diversity of loess deposits across Europe. Based on their physical properties, stratigraphy, or source material areas, these deposits can be divided into characteristic domains (e.g., Lehmkuhl et al., Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021).
Loess of southern Poland is part of the Northern European loess belt (domain II), which preserves a diverse pedo-sedimentary climate record (Fig. 1). This area was strongly influenced by periglacial processes. Loess here often contains features such as ice-wedge pseudomorphs and cryoturbation structures. Southwestern Poland, south of the Odra River, is part of the Western European continental subdomain (IIb). The loess in this area was formed in a relatively narrow corridor between the fluctuating margin of Weichselian Ice Sheet (WIS) in the north and the Sudetes Mountains in the south (Jary, Reference Jary2010; Lehmkuhl et al., Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021). Loess deposition here occurred in short but sometimes very intense phases, interrupted by periods of soil development or erosion. There is increasing evidence pointing to a link between the fluctuations of the WIS front (Fig. 1; Marks, Reference Marks2012) and the dynamics of loess depositional environments in this region (Zöller et al., Reference Zöller, Fischer, Jary, Antoine and Krawczyk2022; Kirsten et al., Reference Kirsten, Starke, Bauriegel, Müller, Jouaux, Lüthgens, Sinapius and Hardt2024).
Figure 1. Locations of key loess–paleosol sections (LPS) located in SW Poland. Labels I, IIb, IIc, and IIe refer to loess subdomains delimited by Lehmkuhl et al. (Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021). Loess covers are in yellow. Glaciation extent (light blue lines) after Marks (Reference Marks2012).
Loess covers in the area were deposited primarily during the last glacial period and are distributed as isolated patches of varying thickness, rarely exceeding 10 meters (Jary, Reference Jary1996, Reference Jary2007, Reference Jary2010; Jary et al., Reference Jary, Kida and Śnihur2002, Reference Bertran, Liard, Sitzia and Tissoux2016). It is assumed that loess was produced mainly by glacial processes and then probably redistributed by the Great Odra Valley fluvial/meltwater system (Smalley et al., Reference Smalley, O’Hara-Dhand, Wint, Machalett, Jary and Jefferson2009; Badura et al., Reference Badura, Jary and Smalley2013), so as to be deposited by northwesterly winds. Those loess deposits transition into the Central European continental subdomain (IIc), which stretches in the area of the Wisła (Vistula) River basin toward western Ukraine (Lehmkuhl et al., Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021). The loess and loess-derived sediments here are primarily found in the southern uplands, such as near Lublin, Sandomierz, and Kraków (Fig. 2). Here, loess deposits can reach thicknesses of 20 m and frequently contain cryoturbation horizons and larger ice-wedge pseudomorphs (Jary, Reference Jary2009; Jary and Ciszek, Reference Jary and Ciszek2013). This characteristic is said to reflect past regional climatic conditions, which were more continental in the east (Cegła, Reference Cegła1972; Jersak, Reference Jersak1973; Maruszczak, Reference Maruszczak and Maruszczak1991; Jary, Reference Jary2007, Reference Jary2009).
Figure 2. (A) Distribution of loess cover in Poland (modified from Jary, Reference Jary2009). (B) Distribution of loess in Europe (Lehmkuhl et al., Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021). (C) Location of the Trzebnica LPS within the Trzebnica Hills. (D) Loess exposure at the site.
The Trzebnica Hills are the northernmost loess region of western Poland. The first mention of silt deposits of aeolian genesis in this area comes from the nineteenth century (Orth, Reference Orth1872). It was not until almost 60 years later that this concept was revisited, and general descriptions of loess deposits in Trzebnica Hills were made (Czajka, Reference Czajka1931; Schwarzbach, Reference Schwarzbach1942; Rokicki, Reference Rokicki1952a, Reference Rokickib). More recent and detailed investigations made by Jary et al. (Reference Jary, Chodak and Krzyszkowski1990) and Jary (Reference Jary and Maruszczak1991) coincided with extensive archaeological work carried out there (Lower Paleolithic site Trzebnica 2) by Burdukiewicz (Reference Burdukiewicz1990, Reference Burdukiewicz1991, Reference Burdukiewicz1993, Reference Burdukiewicz1994). It was then that the existence of ∼6-m-thick silty sediments was once again confirmed for the site, with characteristic solifluction horizons and streaked/laminated loess (Jary et al., Reference Jary, Chodak and Krzyszkowski1990; Jary, Reference Jary and Maruszczak1991). The first, complete stratigraphic description of the site, based on physical and chemical properties of loess, was proposed by Jary (Reference Jary and Maruszczak1991, Reference Jary1996) and by Jary and Krzyszkowski (Reference Jary and Krzyszkowski1994). Those findings showed that the Trzebnica loess was deposited during the last glacial period and is underlain mainly by Neogene clays and/or a locally preserved stony pavement—residuum from Pleistocene-age glacial tills. These descriptions failed to match the stratigraphy proposed by Burdukiewicz (Reference Burdukiewicz1991, Reference Burdukiewicz, Burdukiewicz and Ronen2003), who presented a stratigraphic model based on previous archaeological finds. This model was already highly controversial at the time of publication, due to Burdukiewicz’s hypothesis that the lower part of the loess section, along with artifacts found there, dated to several hundred thousand years ago, namely from the Holsteinian Interglacial (MIS 11).
Determining the chronostratigraphy of loess profiles often relies on independent research methods, such as luminescence dating. Our previous experience in establishing the chronostratigraphy of loess profiles (Moska et al., Reference Moska, Adamiec and Jary2011, Reference Moska, Adamiec and Jary2012, Reference Moska, Jary, Adamiec and Bluszcz2015) based on optically stimulated luminescence (OSL) dating, paired with an appropriate sampling strategy, made it possible to develop detailed chronostratigraphies back to MIS 5 (100 ka). Further detailed luminescence studies for main Polish loess profiles (Moska et al., Reference Moska, Adamiec, Jary and Bluszcz2017, Reference Moska, Adamiec, Jary, Bluszcz, Poręba, Piotrowska, Krawczyk and Skurzyński2018, Reference Moska, Jary, Adamiec and Bluszcz2019a, Reference Moska, Jary, Adamiec and Bluszcz2019b) showed that it is possible to precisely determine sedimentation rates for the youngest loess L1LL1 unit (MIS 2 loess; Marković et al., Reference Marković, Stevens, Kukla, Hambach, Fitzsimmons, Gibbard and Buggle2015) and to compare these rates to data obtained for the other loess profiles in Europe.
The main goal of this paper is to revise the Trzebnica stratigraphies established in earlier works (Burdukiewicz Reference Burdukiewicz1991, Reference Burdukiewicz, Burdukiewicz and Ronen2003; Jary Reference Jary and Maruszczak1991, Reference Jary1996; Jary and Krzyszkowski, Reference Jary and Krzyszkowski1994) based on new OSL and radiocarbon dates. During this revision, we also were able to reconstruct the paleoenvironment of Trzebnica Hills during loess deposition. Together with lithological and geochemical analyses, our findings enabled us to correlate the Trzebnica LPS with nearby LPSs Zaprężyn (Zöller et al., Reference Zöller, Fischer, Jary, Antoine and Krawczyk2022; Jary et al., Reference Jary, Krawczyk, Moska, Piotrkowska, Poręba, Raczyk, Skurzyński, Łopuch and Zöller2023) and Biały Kościół (Moska et al., Reference Moska, Adamiec and Jary2011, Reference Moska, Adamiec and Jary2012, Reference Moska, Jary, Adamiec and Bluszcz2019a).
Study area and geological setting
Loess was collected in 2023 from a newly prepared and documented loess profile at the former Trzebnica brickyard (51°18’44”N, 17°4’14”E, 193 m asl), known since the early twentieth century. The brickyard extracted raw materials and closed in the 1990s. The site is located in the northeastern part of Trzebnica (Fig. 2), on the southern slopes of Winna Góra Mountain (219 m asl). It exists within an archeological site Trzebnica 2, where a documentation site called “Lessy Winnej Góry” (Loess of Winna Góra Mountain) was established in 2016.
This area is part of the Trzebnica Hills, which extend latitudinally for ∼80 km, and are 5–10 km in width. The highest point in this range is the Farna Góra Mountain near Trzebnica, reaching 256 m above sea level. The Neogene and most of the Pleistocene sediments underlying the area appear as deformed structures. These sediments were glaciotectonically deformed—first during the Elsterian (San, MIS 12) glaciation, then during the Saalian (Odra, MIS 6) glaciation—and later acted as a barrier to the Wartanian ice sheet (MIS 6), as confirmed by investigation of the Siedlec Sandur (Krzyszkowski, Reference Krzyszkowski1993). The deformed strata are overlain by a gravel pavement (Krzyszkowski, Reference Krzyszkowski1993). The upper sediments are not deformed. The glaciotectonic origin of the Trzebnica Hills, and by extension of the Trzebnica Rampart, is still a matter of contention (Brodzikowski, Reference Brodzikowski1987).
Late Pleistocene loess at Trzebnica overlies residuum of older Pleistocene sediments or sometimes directly on Neogene clays. The loess cover at the Winna Góra Mountain clay pit is highly variable. Because it lies on an inclined surface (10–30°), the loess at Trzebnica displays structures suggestive of slope redeposition (Jary et al., Reference Jary, Chodak and Krzyszkowski1990; Jary, Reference Jary and Maruszczak1991, Reference Jary1996; Jary and Krzyszkowski, Reference Jary and Krzyszkowski1994). In some places, the contact zone between the Neogene clays and loess is distinct and clear, and the dip of upper surface of the clays follows that of the loess strata and the current land surface (Jary and Krzyszkowski, Reference Jary and Krzyszkowski1994). For the most part, the contact zone contains sandy or gravelly lenses.
At Trzebnica, silty deposits (loess and loess-derived sediments) with a thickness up to 8 meters form the southern vertical wall of the 2023 excavation, which is about 50 meters long (Fig. 2). These deposits rest on a stony pavement of variable thickness (0.0–0.7 m), where the previously mentioned Lower Paleolithic artifacts were discovered (Burdukiewicz, Reference Burdukiewicz1991).
Sampling and methods
Sampling
For OSL analyses, 15 samples from the 8-m-thick loess profile at Trzebnica were collected from the most characteristic profile sections. Samples were taken from a clean vertical section using thin-walled steel pipes from an Eijkelkamp system. Roughly 0.5 kg of material was taken from the area around the sample tubes and placed into plastic bags for gamma spectrometry (dose rate determinations). The OSL ages obtained from the site were correlated to four radiocarbon dates. Our inability to find charcoal fragments for radiocarbon dating necessitated the recovery of about 2 kg of raw loess for each sample for humic acid dating.
The samples for lithological characterization were collected using continuous column sampling (CCS) method at 2-cm intervals (Antoine et al., Reference Antoine, Rousseau, Moine, Kunesch, Hatte, Lang, Tissoux and Zöller2009; Jary et al., Reference Jary, Krawczyk, Moska, Piotrkowska, Poręba, Raczyk, Skurzyński, Łopuch and Zöller2023). This technique relies on the careful cutting of continuous material to create a complete record of lithological variability and thus eliminates gaps in the record.
Chemical pre-treatment and luminescence measurements
Standard chemical pre-treatments were used for OSL measurements (Moska et al., Reference Moska, Bluszcz, Poręba, Tudyka, Adamiec, Szymak and Przybyła2021). Medium-sized quartz grains (45–63 μm) were first treated with 20% hydrochloric acid (HCl) and 20% hydrogen peroxide (H2O2). After that, two solutions of sodium polytungstate of densities 2.62 g/cm3 and 2.75 g/cm3 were used in density separation procedures to obtain pure quartz grains. Finally, the quartz grains were sieved twice, before and after a 40-minute etch with concentrated hydrofluoric acid (HF). After this step, pure quartz grains were ready for luminescence measurements on our Risø TL/OSL DA-20 reader, equipped with a calibrated 90Sr/90Y beta source delivering approximately 6.0 Gy/min to grains at the sample position. It was employed for all OSL measurements, and a 6-mm Hoya U-340 filter was applied.
Dose rate calculations
Prior to dose rate calculations, the samples were dried, placed in the measurement containers, and stored for a minimum of three weeks to ensure radioactive equilibrium in the decay series between gaseous 222Rn and 226Ra in the 238U decay chain. The activities of 238U and 232Th series and 40K for dose rate were determined by means of low-background, high-resolution gamma spectrometry analysis. The measurements were performed on ∼800-g samples for 24 hours. The activities of the isotopes present in the sediment were determined using IAEA standards RGU, RGTh, and RGK after subtraction of the detector background.
Long-term water content was assumed to have been 15 ± 5%, because for this type of sediment we were not able to reproduce precisely its changes for the last several thousand years. Therefore, the accepted value with greater uncertainty lies within the range of water content for this type of sediment in Europe. Dose rates were determined through an online calculator (Tudyka et al., Reference Tudyka, Koruszowic, Osadnik, Adamiec, Moska, Szymak, Bluszcz, Zhang, Kolb and Poręba2023) incorporating the latest conversion factors. Cosmic ray dose rates for the site were computed following Prescott and Stephan (Reference Prescott and Stephan1982). Dose rate data are summarized in Table 1.
Table 1. Relevant data for investigated luminescence samples: sample codes, depth, radionuclide concentration, dose rate, equivalent dose (CAM model), and final age.
Luminescence procedures and De (equivalent dose) calculation
Equivalent doses were determined using the single-aliquot regenerative-dose (SAR) protocol (Murray and Wintle, Reference Murray and Wintle2000) on 6-mm aliquots with a preheat of 260°C for 10 s and a cut-heat of 220°C. SAR dose-response curves were best described by a single saturating exponential function. A 0-Gy regenerative-dose step was included in all SAR sequences to monitor recuperation, which remained below 5% for all aliquots. Recycling ratios were within 10% of unity for every sample, confirming effective correction for sensitivity changes. Dose-recovery tests were performed for six samples using the same SAR conditions. Five aliquots per sample were bleached with blue light for 100 s at room temperature, allowed to rest for 10,000 s, and bleached again for 100 s. After bleaching, a laboratory dose comparable to the equivalent dose for each sample was administered and subsequently measured using the SAR protocol. Recovery ratios (D e/given dose) were close to unity, ranging from 0.9 to 1.0.
Final equivalent dose (D e) values were calculated for all samples using the Central Age Model (CAM) (Galbraith et al., Reference Galbraith, Roberts, Laslett, Yoshida and Olley1999) using R package ‘Luminescence’ (Kreutzer et al., Reference Kreutzer, Burow, Dietze, Fuchs, Schmidt, Fischer, Friedrich, Riedesel, Autzen and Mittelstrass2020). All obtained dose distributions are presented in Figure 3, where the dose distributions are presented in terms of relative probability density functions (Berger, Reference Berger2010). Overdispersion parameters were calculated for all samples and in all cases were much lower than 20%. Dose distributions for all samples look generally unimodal, and thus, the CAM model in our opinion is the most appropriate for final equivalent dose calculations. For all samples, it can be assumed that the sediment represents well-bleached quartz (Moska et al., Reference Moska, Jary, Adamiec and Bluszcz2019a).
Figure 3. Dose distributions presented in terms of relative probability density functions (Berger, Reference Berger2010) for all investigated samples.
Radiocarbon dating
Radiocarbon dates were obtained for four loess samples. Despite an intensive search for the best suitable radiocarbon materials (e.g., charcoal or snail shells), the loess lacked such material. Thus, our best option was to perform radiocarbon analyses on humic acids in the loess. A major challenge in 14C dating of sediments is to extract enough organic matter to provide a reliable radiocarbon age. In the laboratory, all samples were analyzed using an optical microscope with a high-quality digital camera, to perform a preliminary characterization of the organic material. Charcoal fragments were recovered from only one sample; these were chemically treated using NAOH sol protocol. The other soil samples were chemically treated for their humic acid fraction (alkali-soluble fraction, NAOH sol). We recognize the risk that the 14C ages may be affected by humic acids originating in overlying, younger organic material in the soil (Wild et al., Reference Wild, Steier, Fischer and Höflmayer2013).
The radiocarbon age for each sample was determined using the AMS technique. Subsamples of the extracted material were packed into tin capsules, combusted in a Vario Micro Cube Elementar elemental analyzer, and then graphitized using an AGE-3 system (Němec et al., Reference Němec, Wacker and Gäggeler2010, Wacker et al., Reference Wacker, Němec and Bourquin2010a). The 14C concentration was measured using the MICADAS accelerator mass spectrometer at the Gliwice Radiocarbon Laboratory (Wacker et al., Reference Wacker, Bonani, Friedrich, Hajdas, Kromer, Němec, Ruff, Suter, Synal and Vockenhuber2010b; Ustrzycka et al., Reference Ustrzycka, Piotrowska, Kłusek, Pawełczyk, Michczyńska, Michczyński, Kozioł and Jędrzejowski2025). Phthalic Anhydride (Sigma-Aldrich) was used as a background material and Oxalic Acid II (NIST SRM4990C) was used as a modern reference material. In addition to samples of unknown radiocarbon age, background and standard samples, secondary IAEA standards were measured to control the accuracy of the measurements and were analyzed as unknowns (Ustrzycka et al., Reference Ustrzycka, Piotrowska, Kłusek, Pawełczyk, Michczyńska, Michczyński, Kozioł and Jędrzejowski2025). Finally, the radiocarbon age and its uncertainty were determined in BATS software (Wacker et al., Reference Wacker, Bonani, Friedrich, Hajdas, Kromer, Němec, Ruff, Suter, Synal and Vockenhuber2010b) and calibrated using the OxCal program (Bronk Ramsey, Reference Bronk Ramsey2009) and the IntCal20 calibration curve (Reimer et al., Reference Reimer, Austin, Bard, Bayliss, Blackwell, Bronk Ramsey and Butzin2020).
Age–depth model
A key benefit of high-resolution luminescence dating is its ability to leverage statistical software such as Oxcal (Bronk Ramsey, Reference Bronk Ramsey2009; Bronk Ramsey and Lee, Reference Bronk Ramsey and Lee2013) to generate an age–depth model, which represents Bayesian accumulation histories for sedimentary deposits. This model can incorporate radiocarbon and other calendar-scale dates, such as luminescence ages. Oxcal processes a core by dividing it into very thin vertical layers (with a default resolution of 0.5 cm) and using millions of Markov Chain Monte Carlo (MCMC) iterations, calculates the sedimentation rate (in years per centimeter) for each layer.
Grain-size analyses and end-member modeling
Grain size distribution data were determined using a laser diffractometer Mastersizer 2000 (manufactured by Malvern, England). Before the measurement, samples were pretreated with 20% H2O2 and 10% HCl to remove organic matter and carbonates. For better dispersion, sodium hexametaphosphate (Calgon™) was added to the solution before measurement. Grain-size results are presented according to Wentworth’s (Reference Wentworth1922) classification, divided into five fractions: clay (< 4 μm), fine silt (4–16 μm), medium silt (16–32 μm), coarse silt (32–63 μm), and sand (> 63 μm).
To detect transitions from low to high wind dynamics, the U-ratio was calculated for grain size data, as (the grain-size boundaries were retained according to the original scheme):
\begin{equation*}u - ratio = \frac{{16{\text{ }}to{\text{ }}44{\text{ }}\mu m}}{{5,5{\text{ }}to{\text{ }}16{\text{ }}\mu m}}\end{equation*}(Vandenberghe, Reference Vandenberghe1985; Vandenberghe et al., Reference Vandenberghe, Zhisheng, Nugteren, Huayu and Van Huissteden1997; Vandenberghe and Nugteren, Reference Vandenberghe and Nugteren2001). To calculate wind dynamics and atmospheric dust, the grain-size index (GSI) was used (Rousseau et al., Reference Rousseau, Antoine, Hatté, Lang, Zöller, Fontugne and Othman2002, Reference Rousseau, Antoine, Kunesch, Hatté, Rossignol, Packman, Lang and Gauthier2007; Antoine et al., Reference Antoine, Rousseau, Moine, Kunesch, Hatte, Lang, Tissoux and Zöller2009). GSI was calculated as:
\begin{equation*}GSI = {\text{ }}\frac{{26{\text{ }}to{\text{ }}52,6{\text{ }}\mu m}}{{ \lt 26{\text{ }}\mu m}}\end{equation*}Regarding the process-related interpretation of grain-size data (Prins and Vriend, Reference Prins and Vriend2007; Vriend et al., Reference Vriend, Prins, Buylaert, Vandenberghe and Lu2011; Dietze et al., Reference Dietze, Hartmann, Diekmann, IJmker, Lehmkuhl, Opitz, Stauch, Wünnemann and Borchers2012; IJmker et al., Reference IJmker, Stauch, Dietze, Hartmann, Diekmann, Lockot, Opitz, Wünnemann and Lehmkuhl2012; Vandenberghe, Reference Vandenberghe2013), end-member modeling analysis (EMMA) enables the reconstruction of a limited number of genetically meaningful grain-size populations within a dataset. The fundamental principles of this approach were introduced by Weltje (Reference Weltje1997) and further developed by Weltje and Prins (Reference Weltje and Prins2003), with subsequent methodological improvements by Dietze et al. (Reference Dietze, Hartmann, Diekmann, IJmker, Lehmkuhl, Opitz, Stauch, Wünnemann and Borchers2012). More recent developments, including an openly accessible R implementation (“EMMAgeo”, http://CRAN.R-project.org/package=EMMAgeo), were provided by Dietze and Dietze (Reference Dietze and Dietze2013). In the present study, the optimal number of end-members (q) was determined using the parameter test implemented in EMMAgeo (test.parameters()), which evaluates the explained variance (R2) across combinations of q and weight limits (l). To obtain a robust criterion comparable to the classical “elbow test”, we calculated the mean explained variance for each q by averaging R2 across all tested weight limits (Nottebaum et al., Reference Nottebaum, Stauch, Hartmann, Zhang and Lehmkuhl2015). The resulting curve of mean R2 versus the number of end-members showed a clear inflection point at q = 4, where the gain in model performance diminished relative to higher q values (Fig. 4).
Figure 4. Plot of R2 against number of end-members (q); justifying the choice of EM-model with q = 4.
Carbonates and organic matter
Carbonate contents of the samples were determined using the Scheibler method (Scheibler, Reference Scheibler1867). The method involves measuring the volume of carbon dioxide released during the reaction of calcium carbonate present in the loess sample with 10% HCl. The result is calculated based on the ratio of the released CO₂ to the sample mass.
Determination of organic substances in the samples was carried out using the Tyurin method (Tyurin, Reference Tyurin1935). In this direct volumetric method, the humus content is quantified based on the amount of oxidizing agent reduced—in this case, potassium dichromate (K₂Cr₂O₇). The excess of unreduced potassium dichromate is titrated (neutralized) using Mohr’s salt—a standardized solution of ferrous ammonium sulfate.
Results
Grain-size characteristic of Trzebnica LPS
The grain-size fraction distributions exhibit a clear and systematic vertical variability throughout the entire profile, reflecting changing conditions of aeolian deposition, as well as post-depositional modification of the sediments. The clay fraction remains relatively low across the profile, generally below 10%, with localized increases to approximately 12–15% in the middle and lower sections. Maximum clay contents are observed at depths of ca. 5.5–6.5 m (Fig. 5), coinciding with an increase in fine silt proportions and a concomitant decrease in transport-dynamics indicators.
Figure 5. Lithology, granulometry, CaCO3, and C organic contents plotted against lithological units of Trzebnica LPS. The red box represents the CCS location.
Coarse silt constitutes the dominant sediment component over most of the profile, locally exceeding 35–38%, particularly within the depth interval of ∼3.0–5.0 m. The lowest coarse silt contents (<20%) occur in horizons enriched in fine silt and clay fractions. Medium silt displays a more uniform distribution, typically ranging between 15 and 25% and lacking abrupt fluctuations. In contrast, fine silt shows enhanced contributions, reaching up to ∼30–35%, mainly in the uppermost and lowermost parts of the profile (Fig. 5).
Sand content is generally low and does not exceed 5–10% throughout most of the sequence. However, several distinct horizons exhibit pronounced enrichment in this fraction, with values temporarily rising to 15–20%. These episodes are interpreted as short-lived, high-energy depositional events associated with short-range aeolian transport and increased wind competence. The raw grain-size data is provided in Supplement 1.
End-member modeling analyses and grain size indicators
End-Member Modelling Analysis (EMMA) distinguished four statistically robust end members (EM1–EM4) whose relative contributions vary systematically with depth, reflecting changes in sediment supply, transport energy, and post-depositional modification. EM3, corresponding to coarse silt with a modal grain size of ∼63 μm, is the dominant component throughout most of the profile. Its contribution locally reaches very high values of approximately 70–80%, particularly within the middle part of the section, between ∼3.0 and 5.5 m depth (Fig. 6). This interval coincides with pronounced maxima in both the Grain-Size Index (GSI) and the U-ratio, indicating deposition under highly dynamic aeolian conditions characterized by strong winds and efficient deflation of proximal dust sources. In contrast, the lowest proportions of EM3 (<30–40%) occur in depth intervals enriched in finer fractions, where indicators of transport energy systematically decrease.
Figure 6. Results of the EMMA and grain-size indicators for the Trzebnica 2 profile. (A) Vertical variability of the EMMA scores; (B) variability of grain-size indicators: U-ratio, GSI, Mz (mean grain size), and Md (modal grain diameter); (C) grain-size distributions of the various end members; the shares of the total variance associated with each of the four end members is given in the legend.
EM2, representing fine silt with a modal size of ∼16 μm, shows a marked increase in the upper part of the profile and again below approximately 6.0 m depth, where its contribution commonly reaches 40–50% (Fig. 6). These zones correspond to intervals dominated by finer granulometric fractions and reduced values of GSI and U-ratio, suggesting weaker wind regimes, longer transport distances, and more stable depositional conditions. EM1, characterized by very fine silt to clay-sized particles (∼4 μm), remains a secondary component throughout the sequence, generally contributing less than 15–20%. However, its proportion increases distinctly in horizons affected by gleying and enhanced post-depositional alteration, where elevated moisture availability and reduced aeration favored the retention and accumulation of the finest particles.
EM4, associated with sandy material, occurs sporadically and reaches maximum values of about 10–15%. Its discontinuous distribution underscores the episodic nature of coarse-grained input, interpreted as short-lived, high-energy depositional events linked to local or near-source sediment supply. Importantly, the grain-size distributions of all end members (EM1–EM4) exhibit bimodal characteristics. In each case, a secondary, less prominent peak is expressed within the sandy size range, typically around 125 μm and 500 μm. This bimodality represents a clear signal of admixture of redeposited slope or locally derived material, superimposed on the primary aeolian dust populations. Such patterns highlight the complex interplay between long-range aeolian transport and intermittent local sediment reworking that governed the formation of the studied sequence (Fig. 6).
CaCO₃ and organic carbon contents
Calcium carbonate (CaCO₃) content displays pronounced contrasts between different parts of the profile. In the middle section, CaCO₃ values reach maxima of approximately 6–8%, locally exceeding 8%, which corresponds with the dominance of EM3 and high values of GSI and U-ratio. In contrast, the upper and lower parts of the profile show a marked decrease in CaCO₃ content to below 2–3%, and locally close to zero. These low values are interpreted as the result of decalcification processes and prolonged pedogenic alteration (Fig. 5).
Organic carbon (C_org) contents remain low in intervals characterized by intensive aeolian accumulation, generally below 0.2–0.3%. Distinct increases in organic carbon, reaching ∼0.8–1.2% and locally exceeding 2.5%, are observed in selected horizons, mainly in the lower and partly in the middle sections of the profile. These enrichments coincide with increased proportions of fine-grained fractions and reduced grain-size indices. Peaks in C_org frequently correspond to local decreases in CaCO₃ content, indicating episodes of surface stabilization and enhanced biological activity (Fig. 5). The raw organic carbon and CaCO₃ data is provided in Supplement 2.
Characteristics and interpretation of lithological units of the Trzebnica loess–paleosol sequence
Based on laboratory analyses (grain-size distribution, organic carbon content, and calcium carbonate content) combined with detailed field investigations of the Trzebnica loess–paleosol sequence (LPS) conducted in December 2023, 10 lithological units were distinguished from top to bottom (Figs. 5 and 6).
U1—modern brown soil (0.00–0.95 m)
The uppermost unit represents a well-developed soil horizon comprising a dark, sandy A horizon with a crumb structure and an underlying brown Bw horizon (Fig. 7B) containing a substantial admixture of material derived from the A horizon. The organic carbon (C_org) profile reaches its highest values within the entire section (up to ∼2.5%), whereas calcium carbonate (CaCO₃) is nearly absent (maximum ∼0.5%). End-member modeling indicates the dominance of EM2 (modal ∼28 μm) and EM1 (modal ∼8 μm), while both GSI and U-ratio display some of the lowest values recorded in the profile. These characteristics collectively reflect long-term surface stability associated with intensive pedogenic processes and negligible net aeolian silt accumulation. Slight enrichment in sand in the uppermost part of the unit suggests minor modern aeolian input of locally derived material.
Figure 7. Photos of parts of the Trzebnica LPS. (A) Entire Trzebnica section; (B) upper part of section with modern soil and ice-wedge pseudomorph; (C) laminated loess in the middle part of profile; (D) deformation zone; (E) gleyed material at bottom part of profile; (F) contact with Neogene clay.
U2—streaked sediments that also fill a collapsed ice-wedge pseudomorph, modified by thermokarst processes (0.95–1.65 m)
Unit U2 fills a deformed ice-wedge pseudomorph and is characterized by inclined streaks, internal layering, and local sand enrichment, indicating secondary deposition related to thermokarst subsidence (Fig. 7B). The sediments overlying the ice-wedge pseudomorph likely were re-deposited on a hillslope after the erosion of a former ice-wedge pseudomorph by thermokarst processes. Medium- and fine-silt fractions dominate and correspond to EM2; however, increased sand content in the upper part coincides with a local contribution of EM4, most likely reflecting short-distance transport of local material and/or downslope reworking. Low GSI and U-ratio values, along with the absence of CaCO₃, point to weak aeolian dynamics and sedimentary processes associated with permafrost degradation rather than active loess accumulation.
U3—gleyed loess affected by permafrost conditions (1.65–2.70 m)
Unit U3 consists of gleyed loess marked by horizontally elongated gray mottles (Fig. 7B). Sand content increases in the upper part, whereas the lower portion is enriched in coarse silt. Locally, darker intercalations rich in organic carbon occur near the boundary between the material surrounding the ice-wedge fill and the fill itself. EMMA results indicate a predominance of EM2, with localized increases in EM3 suggesting the onset of more dynamic aeolian deposition. Persistently low GSI and U-ratio values—especially in the upper part—reflect relatively moist environmental conditions, corroborated by well-developed gley features.
U4—calcareous laminated loess: main phase of loess accumulation (2.70–5.35 m)
Unit U4 represents a classic, massive loess deposit rich in medium- and coarse-silt fractions, indicating the dominance of dynamic aeolian processes during deposition. Distinct horizontal lamination, particularly well developed in the upper part (Fig. 7C), is accentuated by accumulations of fine sand and iron concretions, pointing to episodic sedimentation under fluctuating wind regimes. The coarse-silt end-member EM3 (modal ∼63 μm) reaches some of its highest values in the profile, unequivocally indicating intense transport under strong deflation and close proximity to dust source areas. High GSI and U-ratio values confirm cold, dry glacigenic conditions typical of pleniglacial stadial phases, marked by limited vegetation cover and highly effective aeolian transport. A clear increase in CaCO₃ reflects sustained input of fresh aeolian material derived from carbonate-rich proglacial surfaces such as sandur plains and exposed fluvio-glacial terraces. The near absence of C_org indicates minimal pedogenic activity. A subtle mid-unit increase in clay content accompanied by slightly higher C_org values likely reflects brief phases of surface stabilization and reduced aeolian dynamics.
U5—calcareous streaked loess (5.35–5.90 m)
Unit U5 is characterized by calcareous, streaked loess exhibiting numerous gley features and post-sedimentary deformation, indicative of periodic moisture influx and subsequent alteration (Fig. 7D). Relative to U4, the deposit shows a clear increase in fine-silt fractions (<16 μm), reflected in higher EM2 contributions, interpreted as enhanced long-distance dust transport under reduced wind energy. Conversely, EM3 values decrease markedly, signaling diminished deflation intensity. Moderate GSI and U-ratio values support a weakening of typical cold–dry glacigenic conditions. Although CaCO₃ contents remain relatively elevated, they are lower than in overlying massive loess, suggesting reduced fresh dust input and partial carbonate redistribution under wetter conditions. A slight increase in C_org points to episodic vegetation development and increased surface stability.
U6—redeposited dark material with slope-related deformation (5.90–6.25 m)
Unit U6 comprises darker, clearly deformed sediments dipping westward, interpreted as redeposited material transported downslope (Fig. 7D). It is distinguished by elevated clay and fine-silt contents, accompanied by notable humus accumulation. EMMA results show enhanced EM2 values, indicating reworking of previously deposited fine-grained material originally accumulated under low-energy aeolian conditions. EM3 values are relatively low, contrasting with units U4 and U5, and pointing to limited fresh coarse-silt input. Intermediate GSI and U-ratio values suggest a transitional environment dominated by reduced deflation and increased slope processes. CaCO₃ content decreases further, while elevated C_org—especially in the upper part—reflects local stabilization and brief vegetation development. Strong gley features in the lower part indicate episodic waterlogging. Overall, U6 records a phase of diminished aeolian activity and enhanced local reworking under more humid microclimatic conditions.
U7—laminated loess with locally deformed streaks (6.25–7.00 m)
Unit U7 consists of distinctly laminated loess with nearly horizontal, dark, iron-rich laminae locally disrupted by deformation (Fig. 7D and E). This microstructural organization reflects slow, stratified dust accumulation interspersed with pedogenic processes. Granulometrically, an increase in coarse silt is indicated by higher EM3 values, suggesting renewed episodes of strong aeolian transport under cold and dry glacigenic conditions. Elevated GSI and U-ratio values support this interpretation. CaCO₃ content increases in the upper part of U7, confirming fresh carbonate-rich dust input, although overall values remain lower than in adjacent units, implying alternation between high-energy deposition and more stable intervals. Organic carbon remains low, except for darker laminae that indicate short-lived stabilization phases.
U8—massive gleyed loess (7.00–7.60 m)
Unit U8 represents massive loess with pronounced gleying, especially in the lower part where it forms near-horizontal streaks (Fig. 7E and F). A distinct increase in coarse-silt content and EM3 values reflects intensified episodic deflation and short transport distances. High GSI and U-ratio values confirm cold, dry, and dynamic glacigenic conditions. Simultaneously, CaCO₃ content rises markedly, indicating an influx of fresh dust derived from carbonate-rich proglacial source areas. Low C_org values indicate sparse vegetation and limited surface stabilization. Mixing with the underlying stony pavement suggests local sediment disturbance, likely linked to periglacial processes. Gleying reflects periodic water stagnation, potentially associated with seasonal thaw of frozen ground.
U9—heterogeneous gravelly pavement with clay–silt matrix (7.60–7.70 m)
Unit U9 is a distinctly heterogeneous gravel–stone horizon with a clay–silt matrix, forming a sharp lithological boundary between overlying aeolian deposits and the underlying Neogene clay (Fig. 7F). Sand and gravel dominate, representing locally derived material deposited under entirely different processes than loess accumulation. EM1-type sandy populations most likely are not a result of aeolian transport. Variable CaCO₃ and C_org contents reflect strong mixing and reworking. This unit marks the basal limit of Quaternary aeolian sedimentation.
U10—Neogene clay basement (7.70–7.80 m)
The lowermost unit comprises massive, grayish-blue Neogene clay forming the lithological basement of the profile (Fig. 7F). Its homogeneous fine-grained texture and absence of aeolian end-members confirm a depositional environment fundamentally different from the overlying loess sequence. The contact with U9 displays signs of mechanical mixing and gleying caused by hydrological contrast, as the impermeable clay promoted water accumulation and redox alterations in the overlying deposits.
OSL and 14C dating for the Trzebnica LPS
Analyzing the luminescence data (Table 1) confirms that the entire profile has a similar content of natural radioactive isotopes. This characteristic produces minimal fluctuations in the dose rates across the entire depth of the section. The dose rate values are typical of loess deposits, falling within the range of 3 Gy/ka. The equivalent dose probability distributions (Berger, Reference Berger2010; Fig. 3) confirm that the loess here was well bleached prior to burial (Moska, Reference Moska2019) with a typical unimodal distribution. These data confirm the credibility of the OSL results.
The use of dense sampling of the loess profile in Trzebnica also allows for an attempt to estimate past sedimentation rates (Fig. 8). In this analysis, we focused on defining two key boundaries between the stratigraphic units, which are justified based on their distinct characteristics. Ultimately, our findings highlight locally conditioned rapid sedimentation of the loess for the L1LL1 unit (U3–U8). The rate here, 2 mm per year (Fig. 8), is faster than in previous studies (Moska et al., Reference Moska, Adamiec, Jary and Bluszcz2017, Reference Moska, Adamiec, Jary, Bluszcz, Poręba, Piotrowska, Krawczyk and Skurzyński2018, Reference Moska, Jary, Adamiec and Bluszcz2019a, Reference Moska, Jary, Adamiec and Bluszcz2019b). Based on the OSL chronology, the estimated time of loess deposition at Trzebnica is from 18.5 to 15.3 ka, which places it at the end of the Upper Pleniglacial. However, additional sources of uncertainty—such as partial bleaching, microdosimetry, and tilting processes—suggest that the effective chronological uncertainty may be as high as 2000–3000 years. Therefore, the duration of loess accumulation should be interpreted as approximate, and multiphase deposition with short intervals cannot be ruled out (Antoine et al., Reference Antoine, Rousseau, Moine, Kunesch, Hatte, Lang, Tissoux and Zöller2009; Vandenberghe, Reference Vandenberghe2013).
Figure 8. OSL and radiocarbon dating results, along with luminescence age distributions for the Trzebnica LPS. The age–depth model was developed using OxCal based exclusively on luminescence data because the radiocarbon results, which are shown in the Lithology column, were excluded due to interpretative uncertainties. Black squares indicate sampling locations for C14. Red dots indicate sampling locations for OSL.
The results of radiocarbon dating are presented in Table 2. The data for the samples at 5.4 m and 6.7 m do not match the OSL chronology. The 14C calculated age of the sample from 5.4-m depth is around 4000 years older than the OSL date. For the sample from 6.7-m depth, two 14C ages were retrieved (after the same chemical preparation), resulting in the following ages: 13677 ± 57 years BP for the first subsample and 23041 ± 57 years BP for the second subsample. The first subsample age is younger, and the second one is older than the OSL date. The calibrated age result from 14C dating of the ice-wedge pseudomorph filling is 12984 ± 59 years BP, whereas the age of the dark lamina separating units U2 and U3 is 6279 ± 33 years BP. The dated material in the ice wedge pseudomorph filling was charcoal, while the other age was from humic acids extracted from the lamina.
Table 2. Radiocarbon results from Trzebnica, before and after calibration.
Discussion
General characteristics of the Trzebnica LPS
The Trzebnica LPS, which is composed mainly of laminated and streaky loess facies and weakly gleied horizons, is characterized by a relatively low diversity of litho-pedological features. Gley horizons in Trzebnica are usually discontinuous, and gley processes manifest as spots and lenses. They are not important for stratigraphic interpretation. These features may result from a weakening of the loess sedimentation rate and an increase in humidity. It is assumed that lamination in loess is mainly the result of sediment redeposition through surface-wash processes caused by rainfall or snowmelt (Cegła, Reference Cegła1972; Mücher and De Ploey, Reference Mücher and De Ploey1977; De Ploey, Reference De Ploey1984). In exceptional cases, lamination may be the result of aeolian processes (e.g., dust storms). Lamination formation in loess was studied experimentally by Cegła (Reference Cegła1972), and the aeolian origin of lamination was confirmed, among others, in the Upper Pleniglacial loess in Nussloch (Antoine et al., Reference Antoine, Rousseau, Moine, Kunesch, Hatte, Lang, Tissoux and Zöller2009). Streaking features may be the result of solifluction, niveo-aeolian deposition, or extremely intense surface wash (Mücher, Reference Mücher1986; Jary, Reference Jary1996). The ice-wedge pseudomorph fill (U2) was most likely redeposited along hillslope due to thermokarst processes. Studies have shown that water tracks transporting water in thermokarst might form along slope, giving way to intense erosion (Toniolo et al., Reference Toniolo, Kodial, Hinzman and Yoshikawa2009; Kanevskiy et al., Reference Kanevskiy, Shur, Jorgenson, Brown, Moskalenko, Brown, Walker, Raynolds and Buchhorn2017). These tracks or gullies promote subsidence of surrounding material and therefore can be rapidly filled (Rowland et al., Reference Rowland, Jones, Altmann, Bryan, Crosby, Geernaert and Hinzman2010). To sum up, loess deposits in Trzebnica show features indicating intense aeolian deposition and synsedimentary redeposition caused by a variety of potential slope processes (Mücher, Reference Mücher1986).
The results of OSL dating (Table 1; Fig. 8) indicate very rapid loess sedimentation (8 m in ca. 3000 years), which occurred at the beginning of the last glacial deglaciation (LGD, Palacios et al., Reference Palacios, Hughes, Garcia-Ruiz and Andres2023) directly before the late glacial. Even during this short interval, the climatic and environmental conditions of the Trzebnica Hills area were apparently not stable. Slope instability here is confirmed by lithological indicators (Figs. 5 and 6), especially the variable content of the coarse-silt fraction, which is a function of the distance from the loess source area and the intensity and strength of the wind (Vandenberghe et al., Reference Vandenberghe, Mücher, Roebroeks and Gemke1985, Reference Burdukiewicz1993; Sun and Ding, Reference Sun and Ding1998; Ding et al., Reference Ding, Sun, Rutter, Rokosh and Liu1999; Schaetzl and Attig, Reference Schaetzl and Attig2013; Vandenberghe, Reference Vandenberghe2013).
The upper part of Trzebnica LPS contains a strongly deformed (most likely due to thermokarst processes) ice wedge pseudomorph. Well-preserved ice wedge pseudomorphs had previously been reported for this site in the 1990s (Jary and Krzyszkowski, Reference Jary and Krzyszkowski1994; Jary, Reference Jary1996). The OSL ages obtained for the ice wedge filling and associated, banded thermokarst deposits (Fig. 8; Table 1) indicate that permafrost degradation occurred in this area later than 15–16 ka ago—directly before or during the onset of the Weichselian Late Glacial.
Radiocarbon dating results (Fig. 8; Table 2), which produced divergent age estimates for the Trzebnica/6.7-m sample, suggest that during chemical treatment there was incomplete separation of humic acids (alkali – soluble) and humin fraction (older sample). Nevertheless, the dates older than 20 ka from the 5.4- and 6.7-m depths confirm that these layers contained organic material, which could have been transported there as a result of slope processes. The 14C cal. age of the charcoal fragments from the fill material in the ice wedge pseudomorph is younger than the age of the surrounding sediment, obtained by OSL dating. The younger age may be due to percolation of younger humic acids from the topsoil into the layer. The significantly underestimated age of the dark lamina at the U2–U3 boundary is the result of contamination by younger organic carbon in the humic acids, again likely because of percolation of humic acids through fissures and large pores in the loess.
The first stratigraphic interpretation of the Trzebnica LPS, proposed by Jary (Reference Jary and Maruszczak1991, Reference Jary1996) and Jary and Krzyszkowski (Reference Jary and Krzyszkowski1994), stated that the Trzebnica loess was deposited mainly during the Upper Pleniweichselian, correlated with MIS 2, with local remnants of gley soils having formed previously, in the Middle Pleniweichselian (MIS 3). This interpretation was a clear contradiction to the stratigraphy proposed by Burdukiewicz (Reference Burdukiewicz1990, Reference Burdukiewicz1991, Reference Burdukiewicz, Burdukiewicz and Ronen2003), who presented a stratigraphic model based on his archaeological research. The model was highly controversial at the time of publication, due to Burdukiewicz’s hypothesis that the lower part of the loess section, where artifacts were found, was formed several hundred thousand years ago, presumably during the Holsteinian Interglacial (MIS 11). Our results indicate that this interpretation is invalid. It is now clear that loess at the Trzebnica site was deposited at the beginning of the LGD.
Because the loess domains distinguished by Lehmkuhl et al. (Reference Lehmkuhl, Nett, Pötter, Schulte, Sprafke, Jary and Antoine2021) are often used in loess regionalization, it is important to point out that the loess of the Trzebnica Hills (Fig. 1) is located on the border of two loess domains: Weichselian marginal or protogenetic zone (Western protogenetic subdomain Ia) and Northern European loess belt (Western European continental subdomain IIb). Therefore, it has characteristics of both of these domains. Lithological and structural features that indicate high dynamics of the loess sedimentary environment are particularly frequent in the loess sequences of the Trzebnica Hills, such as those documented in the Trzebnica LPS.
Updated sedimentary history of the Trzebnica LPS
An important feature of the Trzebnica LPS is the presence of a lithologic discontinuity at its base. In general, lithologic discontinuities are markers of changes in sedimentary history (Schaetzl, Reference Schaetzl1998). Unconformities and discontinuities in the Quaternary sediments of the Trzebnica Hills have already been noted in literature (e.g. Schwarzbach, Reference Schwarzbach1942; Krzyszkowski, Reference Krzyszkowski1993), although their chronostratigraphic positions have not been resolved. At the Trzebnica LPS, an unconformity separates the Upper Pleniweichselian L1LL1 unit (MIS 2) from the underlying Neogene clays, with a locally preserved stony/gravelly pavement in the upper part of the latter (perhaps an erosional residuum associated with Pleistocene tills). There is little doubt that the discontinuity surface in Trzebnica was created as a result of numerous erosion episodes throughout the Pleistocene. The last (youngest) event took place in the period immediately preceding loess sedimentation in Trzebnica and can be correlated with the last glacial maximum (LGM). An unconformity of the same age was recently observed in the Zaprężyn LPS and mentioned by Zöller et al. (Reference Zöller, Fischer, Jary, Antoine and Krawczyk2022) and Jary et al. (Reference Jary, Krawczyk, Moska, Piotrkowska, Poręba, Raczyk, Skurzyński, Łopuch and Zöller2023).
A loess sedimentation phase occurred after a period of strong erosion/deflation (hence the unconformity), which was most likely related to the harsh climatic conditions of the LGM (Hughes and Gibbard, Reference Hughes and Gibbard2015; Palacios et al., Reference Palacios, Hughes, Garcia-Ruiz and Andres2022) and the proximity of the maximum extent of WIS (Jary, Reference Jary1996; Jary and Kida, Reference Jary and Kida2000) during the Leszno and Poznań phases (ca. 24–19 ka; Fig. 1, Marks, Reference Marks2012). During this period, environmental conditions were not conducive to the accumulation of loess in the Trzebnica Hills, while such conditions prevailed in the Niemcza–Strzelin Hills, 70 km to the south (Fig. 1).
Favorable conditions for the accumulation of aeolian silts in the Trzebnica Hills appeared only during the retreat of the Poznań phase ca. 18 ka when loess deposition in the Niemcza–Strzelin Hills was coming to an end (Fig. 9). The process of loess sedimentation in the Trzebnica Hills was intense and short-lived, lasting only about 3 ka. It cannot be ruled out that the sedimentation was briefly interrupted, despite our inability to find clear signs of such episodes. The exceptionally high rate of loess sedimentation in the Trzebnica profile is most likely related to the orography of this area. Loess was deposited on an inclined surface on the leeward side of the local hill, as was deducted from the lithological indices.
Figure 9. Diversity of geomorphic processes in two loess sedimentary zones: the Trzebnica Hills and Niemcza–Strzelin Hills. Phases of the last glacial deglaciation according to Marks (Reference Marks2012).
Unfortunately, we do not have enough evidence to precisely determine the direction of the winds carrying the dust. We think that the loess was deposited in the wind shadow zone on the southern morphological edge of the Trzebnica Hills. The main source areas were located in the north and included extensive glacial till plains and alluvial deposits from ice-marginal valleys. The loess dust was most likely entrained and carried away by northerly and north-easterly katabatic winds during the early phases of the WIS deglaciation, which is consistent with the results published by Schaffernicht et al. (Reference Schaffernicht, Ludwig and Shao2020). Similar conclusions about the role of katabatic winds in the supply of loess dust have recently been reached by Kirsten et al. (Reference Kirsten, Starke, Bauriegel, Müller, Jouaux, Lüthgens, Sinapius and Hardt2024) in northeastern Germany, and by Hedeving et al. (Reference Hedeving, Ekström, Johnson, Alexanderson, Baykal and Stevens2024) in southwestern Sweden.
There is no direct evidence of the existence of continuous permafrost during the loess sedimentation in Trzebnica. Ice wedge pseudomorphs, which are the best evidence of former permafrost, occur only in the upper part of the loess sequence and are strongly deformed by thermokarst processes. The Trzebnica loess was deposited under relatively dry periglacial conditions. Later, an increase in humidity caused the formation of ice wedge structures (LGD). Those structures then faced degradation during onset of the late glacial, caused by an increase in temperature.
Loess sedimentation in SW Poland
Late Pleistocene sedimentation of loess across southwestern Poland occurred in two main phases, both of which are well documented for the Zaprężyn LPS (Trzebnica Hills; Zöller et al., Reference Zöller, Fischer, Jary, Antoine and Krawczyk2022; Jary et al., Reference Jary, Krawczyk, Moska, Piotrkowska, Poręba, Raczyk, Skurzyński, Łopuch and Zöller2023) and Biały Kościół LPS (Niemcza-Strzelin Hills; Moska et al., Reference Moska, Adamiec and Jary2011, Reference Moska, Adamiec and Jary2012, Reference Moska, Jary, Adamiec and Bluszcz2019a; Krawczyk et al., Reference Krawczyk, Ryzner, Skurzyński and Jary2017). At both sites the first sedimentation phase is marked by the L1LL2 loess unit, which is correlated with Lower Pleniweichselian (MIS 4). Although comprised of loess, the unit often has been modified by gelifluction and related slope processes, which were synchronous with or immediately followed dust deposition. The second, younger phase, is correlated with the Upper Pleniweichselian (MIS 2). Intense aeolian processes occurred at that time, which resulted in sedimentation of the youngest litho-pedostratigraphic unit L1LL1.
Despite the short distance between Zaprężyn and Biały Kościół, their respective profiles differ significantly, which may be evidence of dynamic changes in the loess sedimentation environment. These differences were presented by Jary (Reference Jary1996) and Jary and Kida (Reference Jary and Kida2000) in the first studies on the characteristics of the sedimentation environments for loess in southwestern Poland. When reconstructing the loess-forming processes in SW Poland, three zones were distinguished: the main source area, the transition zone, and a main accumulation area. The locations of these zones changed over time and space. The conclusions drawn from the litho-pedostratigraphic analysis of the sequences in Biały Kościół and Zaprężyn seem to confirm these assumptions.
Comparing the Trzebnica Hills sites and Biały Kościół
When comparing the Zaprężyn and Trzebnica LPS with the Biały Kościół LPS, clear differences are evident. The climatic conditions of those areas were significantly different, and the interval of loess sedimentation was not synchronous. While the Trzebnica Hills were in the dynamic protogenetic zone, in which deflation and transport of material dominated, conditions in the Niemcza–Strzelin Hills were more favorable for loess accumulation. Chronostratigraphic data from the Biały Kościół LPS (Moska et al., Reference Moska, Adamiec and Jary2011, Reference Moska, Adamiec and Jary2012, Reference Moska, Jary, Adamiec and Bluszcz2019a) indicate continuous loess sedimentation, terminating at ca. 18 ka. Notably, loess sedimentation began at this same time in the Trzebnica Hills. This concurrence suggests that the loess covers in this part of Poland formed primarily during recession of the glaciation marginal zone to the north. The various processes that occurred in Trzebnica Hills and Niemcza–Strzelin Hills are summarized in Figure 9.
Conclusions
The Trzebnica loess–paleosol sequence (LPS) records a brief but very dynamic phase of aeolian deposition during the Late Pleniglacial. The accumulation of 8 meters of loess within a ca. 3-ka interval reflects high sedimentation rates under periglacial conditions. Lithological indicators, including grain size end-member modeling and granulometric indices, reveal frequent synsedimentary redeposition on a slope surface. The ice-wedge pseudomorph, present in the upper part of the sequence and deformed by thermokarst processes, is evidence of past permafrost existence and its subsequent degradation at the onset of the late glacial. The unconformity at the base of the loess sequence marks a regional erosional surface, likely a result of numerous erosion episodes throughout the Pleistocene. The youngest erosional episode took place during the last glacial maximum (LGM). OSL data from the ice wedge fill and banded thermokarst deposits indicate that permafrost degradation occurred in this area ca. 15 ka ago, directly before or at the onset of the late glacial.
The latest results of research on loess profiles in Trzebnica, Zaprężyn, and Biały Kościół indicate a strong meridional climatic gradient during the LGD. During the LGM, the Trzebnica Hills (Trzebnica and Zaprężyn LPSs) were located within the protogenetic zone, which was dominated by deflation and erosion processes. The short period of sedimentation of the Trzebnica loess began with migration of the southern border of the protogenetic zone northwards during the recession of the WIS.
This study challenges earlier stratigraphic interpretations of the Trzebnica LPS, which were derived from archeological research (Burdukiewicz, Reference Burdukiewicz1991, Reference Burdukiewicz, Burdukiewicz and Ronen2003). Instead of a few hundred thousand years, as proposed in the stratigraphical interpretation by Burdukiewicz (Reference Burdukiewicz1993), our findings show that the loess at the Trzebnica site was deposited over a very short phase of sedimentation (between ca. 18.5 and 15 ka), synchronous with the end of the last glacial maximum (MIS 2).
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/qua.2026.10081.
Author contributions
Agnieszka Szymak: conceptualization, funding acquisition, investigation, supervision, writing of initial draft. Zdzisław Jary: conceptualization, funding acquisition, investigation, supervision, writing of initial draft, post-review writing and editing. Piotr Moska: conceptualization, investigation, supervision, writing of initial draft, methodology. Marcin Krawczyk: investigation, writing of initial draft, laboratory analyses, preparation of figures, methodology, post-review writing and editing. Grzegorz Poręba: investigation, laboratory analyses, preparation of figures. Zuzanna Sowińska: investigation, writing of initial draft, post-review writing and editing. Grzegorz Adamiec: methodology. Michał Łopuch: investigation, preparation of figures. Jerzy Raczyk: investigation, laboratory analyses. Jacek Skurzyński: investigation. Alicja Ustrzycka: laboratory analyses, methodology, post-review writing and editing. Andrzej Wiśniewski: investigation. Andrzej Wojtalak: investigation, laboratory analyses.
Acknowledgments
We would like to thank the reviewers and the guest editor for their valuable comments, which significantly improved our work. The research was performed under the National Science Centre projects No. 2021/41/N/ST10/00169 and 2017/27/B/ST10/01854.
Open access
