Introduction
Since its development, radiocarbon dating has been widely applied in studying geological sites, including fluvial environments. Due to the dynamic nature of fluvial systems, no continuous sedimentation sites are capable of recording long-term environmental changes over thousands of years. However, radiocarbon data from geological profiles have been published for decades, enabling collective analyses that explore links between environmental changes, human influence, and fluvial dynamics. European studies include Benito et al. (Reference Benito, Sopeña, Sánchez-Moya, Machado and Pérez-González2003), Thorndycraft and Benito (Reference Thorndycraft and Benito2006) in Spain, Macklin et al. (Reference Macklin, Benito, Gregory, Johnstone, Lewin, Michczyńska, Soja, Starkel and Thorndycraft2006) in Great Britain, Spain, and Poland, Starkel et al. (Reference Starkel, Soja and Michczyńska2006) in Poland, and Hoffmann et al. (Reference Hoffmann, Lang and Dikau2008) in Germany, which examined river valley evolution during the Late Glacial and Holocene. A broader perspective is provided by Starkel et al. (Reference Starkel, Michczyńska, Gębica, Kiss, Panin and Perşoiu2015), which analyzed climate fluctuations in Central-Eastern European fluvial systems from 60–8 ka cal BP, though without summed probability density distributions (PDFs) of radiocarbon dates.
The authors have investigated climate signals recorded in probability density functions (PDFs) of calibrated radiocarbon dates from Poland. In Michczyńska et al. (Reference Michczyńska, Dzieduszyńska, Petera-Zganiacz, Wachecka-Kotkowska, Krzyszkowski, Wieczorek, Ludwikowska-Kędzia, Gębica and Starkel2022), they analyzed 488 radiocarbon and 179 luminescence dates from sites south of the Weichselian Glaciation’s maximum extent (LGM), from various sedimentary environments. The study found that the number of warming-cooling cycles recorded in sediments matched ice-core records, with radiocarbon date peaks aligning with interstadials and luminescence date peaks with stadials.
Further analysis focused solely on fluvial data from the Łódź region (Central Poland). Dzieduszyńska et al. (Reference Dzieduszyńska, Michczyńska, Petera-Zganiacz, Wachecka-Kotkowska, Wieczorek and Krzyszkowski2023) examined 181 radiocarbon and 38 luminescence dates (50–11.7 cal kBP), showing that major aggradation occurred during MIS2, while radiocarbon date clusters corresponded to GI-5.1, GI-4, GI-3, and GI-1 (including GI-1b and GI-1a/GS-1 transitions).
Building on these findings, Michczyńska et al. (Reference Michczyńska, Dzieduszyńska, Gębica, Krzyszkowski, Ludwikowska-Kędzia, Petera-Zganiacz, Wachecka-Kotkowska and Wieczorek2024) expanded the study to fluvial data from Poland’s Łódź region, Holy Cross Mountains, Subcarpathian Basins, and Carpathians. PDFs were analyzed against INTIMATE stratigraphy (Rasmussen et al. Reference Rasmussen, Bigler, Blockley, Blunier, Buchardt, Clausen, Cvijanovic, Dahl-Jensen, Johnsen, Fischer, Gkinis, Guillevic, Hoek, Lowe, Pedro, Popp, Seierstad, Steffensen, Svensson, Vallelonga, Vinther, Walker, Wheatley and Wistrup2014), showing that radiocarbon date accumulations reflect GI transitions, while luminescence dates align with stadials.
The present research aims to explore multi-scale relationships between regional geomorphological processes and global climate trends (50–15 cal kBP). To achieve this, the authors revisited fluvial data from southern Poland (Starkel et al. Reference Starkel, Soja and Michczyńska2006; Gębica et al. Reference Gębica, Michczyńska and Starkel2015), refining previous analyses by incorporating subgrouping methods. Additionally, to test the approach in different regions, data from eastern Netherlands (Dinkel River valley; van Huissteden Reference Van Huissteden1990) and eastern Germany (Lausitz region and Leipzig area; Cepek Reference Cepek1965; Kohl & Quitta Reference Kohl and Quitta1966; Mol Reference Mol1997; Hiller et al. Reference Hiller, Junge, Geyh, Krbetschek and Kremenetski2004) were included. These new areas were chosen due to the availability of published radiocarbon datasets. The database used in the current study is included in the supplementary materials.
Methods
The analyses are based on a database. For over a decade, the authors have collected radiocarbon and luminescence age determinations from sites across Poland. The database is continuously updated and verified. One example of verification is the Stubno site (Sandomierz Basin), where a previous age determination of Gd-4479 15,200±500 BP (Klimek et al. Reference Klimek, Łanczont, Bałaga and Łanczont1997) was excluded from current analyses because palynological studies and new radiocarbon age determinations showed that the earlier date was overestimated (Kołaczek et al. Reference Kołaczek, Gałka, Apolinarska, Gębica, Superson, Michno, Harmata, Szczepanek, Płóciennik, Gąsiorowski and Karpińska-Kołaczek2017; Gębica et al. Reference Gębica, Michno, Sobucki, Wacnik and Superson2022).
Currently, the fluvial environment database includes 273 dates from sites in the Subcarpathian Basin river valleys and 78 dates from river valleys in the Carpathians (Poland). The database also includes 66 dates for sites in southeastern Germany and 87 dates for sites from the Dutch section of the Dinkel River valley.
The analyses involve constructing summed probability density functions (PDFs) of calibrated radiocarbon dates. This is done using the SUM command in OxCal v.4 software (Bronk Ramsey Reference Bronk Ramsey2009) and the latest calibration curve, IntCal20 (Reimer et al. Reference Reimer, Austin, Bard, Bayliss, Blackwell, Bronk Ramsey, Butzin, Cheng, Edwards, Friedrich, Grootes, Guilderson, Hajdas, Heaton, Hogg, Hughen, Kromer, Manning, Muscheler, Palmer, Pearson, van der Plicht, Reimer, Richards, Scott, Southon, Turney, Wacker, Adolphi, Büntgen, Capano, Fahrni, Fogtmann-Schulz, Friedrich, Köhler, Kudsk, Miyake, Olsen, Reinig, Sakamoto, Sookdeo and Talamo2020).
Some of the collected dates have asymmetric uncertainty values. To perform calibration that takes this information into account, it was necessary to convert them into 14C concentration expressed as F14C. This allowed the results to be obtained on a new scale with symmetric uncertainty values. This approach was necessary because the OxCal software does not support calibration of dates with asymmetrical uncertainties.
Examples of studies focused on the reconstruction of fluvial environments based on probability distributions of calibrated radiocarbon dates are well known in the literature (e.g., articles cited in the Introduction). However, the authors believe that an additional value may lie in presenting such distributions in the context of the INTIMATE stratigraphy (Rasmussen et al. Reference Rasmussen, Bigler, Blockley, Blunier, Buchardt, Clausen, Cvijanovic, Dahl-Jensen, Johnsen, Fischer, Gkinis, Guillevic, Hoek, Lowe, Pedro, Popp, Seierstad, Steffensen, Svensson, Vallelonga, Vinther, Walker, Wheatley and Wistrup2014). This approach may reveal climatic change trends recorded in fluvial environments, even if individual sites provide only fragmentary (discontinuous) records. These reconstructions are undoubtedly influenced by the preferential selection of samples for dating, which in this case supports such analyses. Limited funding for scientific research often leads to selecting samples for dating from locations where lithological changes are visible in the profile, such as transitions between organic and mineral sediments.
It is worth noting that if changes were synchronous across multiple locations, peaks will appear in the PDF distributions. In contrast, signals of a local nature will be smoothed in the summed analysis. Therefore, summed probability density analysis helps detect supra-local changes that can be compared to global climate changes.
Research areas
Figure 1 illustrates the location of river valleys from which radiocarbon age determinations were analyzed in this study. All these areas are located outside the maximum extent of the Weichselian Glaciation (LGM).
Figure 1. Study area. The bold red segments schematically indicate sections of river valleys with sites from which radiocarbon dates analyzed in this study were collected. The maximum extent of the Weichselian ice sheet at 40, 35, and 30 ka, as well as the LGM, is shown based on Batchelor et al. (Reference Batchelor, Margold, Krapp, Murton, Dalton, Gibbard, Stokes, Murton and Manica2019).
Eastern Netherlands, Dinkel River valley
The Dinkel River valley, described by van Huissteden (Reference Van Huissteden1990), represents an example of tundra river systems during the Last Glacial Period (Weichselian Glacial Stage). The Dinkel River is 89 km long, with a catchment area of 618 km2.
In this valley, a lower floodplain terrace (Late Weichselian–Holocene) and a higher overbank terrace (Weichselian) described as the “coversand plain” were distinguished.
The geological structure of fluvial deposits was identified through drilling, which also allowed for sampling for laboratory analysis and radiocarbon dating. During the Weichselian Glaciation, deposits classified as the Twente Formation (van der Hammen and Wijmstra Reference Van der Hammen and Wijmstra1971) were formed. Several Pleniglacial members of fluvial and aeolian sediments were identified within the Twente Formation, dating from the Early Pleniglacial to the Late Pleniglacial. These were separated by 2–3 erosional surfaces or sedimentation hiatuses. Exposures at the Hengelo site revealed levels of ice wedge casts and cryoturbations (van Huissteden Reference Van Huissteden1990). Fluvial deposits are primarily overbank sediments, while channel deposits cover a much smaller area in the valley floor.
Layers of sandy deposits documenting crevasse splays and natural levees alternate with silty sediments (muddy deposits) and peat deposited in oxbows and thermokarst lakes. Structural and grain size analyses of the deposits indicate the presence of both anastomosing and meandering river systems (van Huissteden Reference Van Huissteden1990).
The main portion of the sedimentary sequence (Twente Formation) consists of sands and silts with interbedded organic layers from the Middle Pleniglacial. Radiocarbon dating of these organic sediments revealed a significant accumulation of radiocarbon dates between 42–40 ka BP, followed by a sharp decline in dates during 39–37 ka BP (van Huissteden Reference Van Huissteden1990). The reduced number of dates in the 39–37 ka BP period was attributed to fluvial erosion and the degradation of permafrost, while the accumulation of dates between 42–40 ka BP was linked to floodplain stabilization and peat formation (van Huissteden Reference Van Huissteden1990).
It is worth noting that in his doctoral thesis, van Huissteden (Reference Van Huissteden1990) analyzed radiocarbon dates from the Dinkel River valley only on the radiocarbon timescale. He also explored correlations between accumulations of radiocarbon dates and the chronostratigraphic divisions of the Middle Pleniglacial. He concluded that “fluctuations in the formation of peat have occurred during the Middle Pleniglacial. However, these fluctuations do not coincide with the classical Middle Pleniglacial interstadials (Denekamp, Hengelo, and Moershoofd).”
A renewed comprehensive analysis of radiocarbon dates after calibration (which van Huissteden could not perform in Reference Van Huissteden1990) and their presentation against the background of INTIMATE stratigraphy (Rasmussen et al. Reference Rasmussen, Bigler, Blockley, Blunier, Buchardt, Clausen, Cvijanovic, Dahl-Jensen, Johnsen, Fischer, Gkinis, Guillevic, Hoek, Lowe, Pedro, Popp, Seierstad, Steffensen, Svensson, Vallelonga, Vinther, Walker, Wheatley and Wistrup2014) would therefore be of interest.
Eastern Germany, Lausitz region and Leipzig area
In eastern Germany (Lausitz region), previous studies examined fluvial formations from the Weichselian Glaciation that filled the Lausitz (Breslau-Magdeburg) ice-marginal valley (Cepek Reference Cepek1965). These formations consist of fine- to medium-grained sands with horizontal, inclined, and trough stratification, and horizontally stratified silts, which in some places show disturbed structures. Periglacial structures, such as ice wedge pseudomorphs and ground convolutions, were also observed in these formations, indicating the presence of permafrost and generally cold, dry climate conditions.
The fluvial formations also contain peats, muddy peats, organic sands, and silts, with occasional fragments of wood and plant remains. Organic deposits from these formations were studied using radiocarbon dating (from 11 to 33 kBP) and palynology, and they were correlated with the Paudorf and Brörup Interstadials (Cepek Reference Cepek1965; Kohl and Quitta Reference Kohl and Quitta1966). Based on this correlation, the age of the fluvial formations was assigned to the Middle and Upper Pleniglacial of the Weichselian Glaciation and the Late Glacial period.
Two phases of fluvial erosion were observed between the deposition of the organic formations associated with the Brörup Interstadial (radiocarbon dates >40 kBP) and the Paudorf Interstadial, (Cepek Reference Cepek1965). Cryoturbation processes were documented in sediments lying between these erosional surfaces. For sediments younger than the Paudorf Interstadial, two phases of fluvial erosion and two phases of cryoturbation development were recorded.
Later studies confirmed these findings but added new details (Hiller et al. Reference Hiller, Junge, Geyh, Krbetschek and Kremenetski2004). Three organogenic complexes were identified in fluvial deposits. The oldest complex recorded evidence of open forest vegetation dominated by Pinus, Betula, Picea, Alnus, Larix, and Cyperaceae, suggesting it may originate from the Early Weichselian or the early phase of the Middle Pleniglacial. This complex is overlain by fluvial sands and gravels that partially eroded it. The middle complex contains two organic sublevels separated by sands with evidence of erosion. The older sublevel was affected by gravitational processes, while the top of the younger sublevel showed signs of erosion in many locations. The age of this complex was estimated to be from approximately at 14C age 37 to >47 kBP, corresponding to the Middle Pleniglacial (Hiller et al. Reference Hiller, Junge, Geyh, Krbetschek and Kremenetski2004). Palynological studies of the middle complex showed a significant increase in NAP (non-arboreal pollen), indicating unforested, open vegetation resembling grass steppe with tundra elements. Organic deposits in this complex likely formed during the Oerel–Moershoofd Interstadials. The youngest organic complex recorded high concentrations of grass pollen and minimal arboreal pollen (Betula, Salix). Radiocarbon dates suggest a possible correlation with the Denekamp Interstadial or the Hengelo–Denekamp Interstadials. According to Hiller et al. (Reference Hiller, Junge, Geyh, Krbetschek and Kremenetski2004), the possibility of localized clusters of open birch-pine forests in floodplains cannot be excluded.
Mol (Reference Mol1997) conducted research on fluvial formations in the Leipzig area and Lausitz (Niederlausitz) region in eastern Germany. In the Leipzig area, three fluvial sedimentary sequences were documented, the youngest of which dates to Holocene. The lowest sequence, associated with braided rivers, can be divided into a portion formed during the Middle Pleniglacial and another formed during the Upper Pleniglacial. Evidence of erosion was observed at the transition between the Middle and Upper Pleniglacial. During the Late Glacial, a middle sequence of alluvium associated with meandering rivers was deposited. Fluvial erosion occurred at the beginning and end of the Late Glacial.
In the Lausitz region, evidence of fluvial processes extends further back in time. A phase of channel incision occurred at the transition between the Lower and Middle Pleniglacial. During the Middle Pleniglacial, rivers exhibited an ephemeral anastomosing character. Another phase of erosion occurred at the transition between the Middle and Upper Pleniglacial. During the Upper Pleniglacial, braided river systems dominated, accompanied by aeolian processes in river valleys.
Fluvial sediments contained interbedded organic deposits, which were radiocarbon dated to between 10 and 48 14C kBP (Mol Reference Mol1997).
A similar approach of examining the summed distributions of calibrated radiocarbon dates within the INTIMATE framework (Rasmussen et al. Reference Rasmussen, Bigler, Blockley, Blunier, Buchardt, Clausen, Cvijanovic, Dahl-Jensen, Johnsen, Fischer, Gkinis, Guillevic, Hoek, Lowe, Pedro, Popp, Seierstad, Steffensen, Svensson, Vallelonga, Vinther, Walker, Wheatley and Wistrup2014) would be beneficial for data from eastern Germany as well.
Southern Poland, Subcarpathian Basins and Carpathians
The history of fluvial environmental changes in the Subcarpathian Basins between 60–8 cal kBP was broadly outlined in the study by Gębica et al. (Reference Gębica, Michczyńska and Starkel2015). Three main river terrace levels were identified in this region, situated at heights of 15–20, 8–12, and 6–8 m above the current river channels. The highest terraces, covered with loess, dominate the western part of the Sandomierz Basin (Upper Vistula Valley) and its southern margin at the outlets of river valleys from the Carpathians. The 15–20 meter terrace contains 2–3 alluvial fills, documenting the rhythm of frequent Interpleniglacial fluctuations (Starkel Reference Starkel2014). The period between 60–25 cal kBP was characterized by alternating processes of alluvial sedimentation and erosional phases. The most intense erosion occurred during the progressive aridification of the climate in the advancing phase of the main stadial of the Last Scandinavian Glaciation (Weichselian Glaciation) between 29–24 cal kBP, when the river cover of this terrace began to be simultaneously incised and overlain with loess. A younger alluvial fill, capped with sands deposited by braided rivers and dated to 20–15 ka BP, was subsequently superimposed. The plain of this 10–12-m terrace was incised in the valleys of the Wisłoka, Wisłok, and San Rivers (Gębica Reference Gębica2004; Starkel Reference Starkel2014; Gębica et al. Reference Gębica, Michczyńska and Starkel2015).
A slightly lower plain, 2–3 m beneath (6–8 m above current river channels), is marked by wide paleomeander bends, mostly active from the Bølling to the Younger Dryas (Kalicki Reference Kalicki and Starkel1991; Starkel Reference Starkel, Starkel, Gregory and Thornes1991, Reference Starkel2003; Szumański Reference Szumański1983). During the Late Glacial period, the river channels transitioned from braided to meandering. In the Younger Dryas, an increase in sediment transport was observed, along with a return to braided channels. The early Holocene, characterized by significant climate changes and forest expansion, resulted in reduced river sediment load and fluctuation of discharges, a shift to meandering channels, and a decrease in meander size.
Gębica et al. (Reference Gębica, Michczyńska and Starkel2015) published a database of radiocarbon and luminescence dates. This data was later used to construct a summed probability density distribution presented in Michczyńska et al. (Reference Michczyńska, Dzieduszyńska, Gębica, Krzyszkowski, Ludwikowska-Kędzia, Petera-Zganiacz, Wachecka-Kotkowska and Wieczorek2024). For the purposes of this study, the database was revised and supplemented with new age determinations published in Kołaczek et al. (Reference Kołaczek, Gałka, Apolinarska, Gębica, Superson, Michno, Harmata, Szczepanek, Płóciennik, Gąsiorowski and Karpińska-Kołaczek2017) and Gębica et al. (Reference Gębica, Michno, Sobucki, Wacnik and Superson2022).
In the Carpathians, the river valleys do not exhibit as distinct terrace levels from the Weichselian Glaciation as in the Subcarpathian Basins. Carpathian terraces typically have erosional bases (strath terraces) overlain by fluvial deposits. In sedimentary series up to 10–15 m thick, representing the entire Pleniglacial period (75–15 kBP), two gravel layers interbedded with slope (solifluction) deposits are sometimes observed. Gravel layers were deposited during rapid flooding events, while solifluction deposits likely formed during permafrost thawing (Starkel et al. Reference Starkel, Michczyńska and Gębica2017). Interbedded within the alluvium are overbank deposits with peat layers dated to the Interpleniglacial period, including an example from Dobra near Limanowa, dated to 32.5 ka BP (Klimaszewski Reference Klimaszewski1971). In the upper layers, fluvial gravel covers are often replaced by slope deposits or represented by alluvial fan deposits from tributaries (Łanczont Reference Łanczont2001). On the foothills of the Carpathians, loess often caps the alluvial series (Gerlach et al. Reference Gerlach, Krysowska-Iwaszkiewicz, Szczepanek and Aleksandrowicz1991; Łanczont Reference Łanczont1993).
Broadly summarizing the history of this region, the Carpathians during Pleistocene glaciations, including the most recent one, were intensely shaped by denudational, fluvial, and aeolian processes. During cold periods, accumulation in river valleys dominated, leading to the formation of alluvial plains. During warmer periods, with increased vegetation, erosion prevailed, and rivers tended to change from braided to meandering. During the early Weichselian Glaciation (corresponding to Marine Isotope Stage 5d-a, MIS 5d-a, Lisiecki and Raymo Reference Lisiecki and Raymo2005), the Carpathians valleys were characterized by wide erosional plains, and rivers incised into older alluvial covers. During the Lower Pleniglacial (MIS 4), intense sediment accumulation occurred, linked to deteriorating climatic conditions. The Middle Pleniglacial (MIS 3) featured alternating accumulation and erosion, with fluvial deposits often interfingering with slope material (e.g., solifluction, colluvial). In the Upper Pleniglacial (MIS 2), phases of deep river incision and new alluvial covers, often capped by loess, were recorded, reflecting a more continental climate.
Given this broad overview of the history of the Subcarpathian Basins and the Carpathians, it is worth revisiting and attempting to reconstruct these events with greater resolution. This is particularly relevant because radiocarbon and luminescence dates from the Interpleniglacial period (58–32 cal kBP) from several selected sites in the Subcarpathian Basins and the Carpathians were analyzed in Starkel et al. (Reference Starkel, Michczyńska and Gębica2017). The study highlighted those frequent cycles of cooling and warming during the Interpleniglacial period significantly influenced periglacial geomorphological processes, including the development of slope covers and fluvial deposits. Earlier studies attempted to correlate organic sediment dates solely with the Denekamp and Hengelo Interstadials, leading to the rejection of some radiocarbon dates or incorrect conclusions. Starkel et al. (Reference Starkel, Michczyńska and Gębica2017) suggested that future analyses should use the INTIMATE stratigraphy framework and consider a greater number of warming and cooling episodes.
In the present study, we aim to follow the suggestions of Starkel et al. (Reference Starkel, Michczyńska and Gębica2017) and conduct a detailed analysis of the frequency distributions of calibrated radiocarbon dates from the Subcarpathian Basins and the Carpathians within the context of INTIMATE stratigraphy.
Results
The PDF curves for the analyzed regions are presented in Figure 2. These curves are shown against the background of INTIMATE stratigraphy, which classifies periods into glacial stadials (GS) and glacial interstadials (GI) based on variations in the δ18O and calcium ion concentration curves from Greenland ice cores (e.g., the North Greenland Ice Core Project – NGRIP).
Figure 2. PDF curves for the analyzed regions. A – Eastern Netherlands, B – Eastern Germany, C – S Poland, Subcarpathian Basins, D – S Poland, Carpathians, E – δ18O curve from the NGRIP ice core and INTIMATE stratigraphic division (Rasmussen et al. Reference Rasmussen, Bigler, Blockley, Blunier, Buchardt, Clausen, Cvijanovic, Dahl-Jensen, Johnsen, Fischer, Gkinis, Guillevic, Hoek, Lowe, Pedro, Popp, Seierstad, Steffensen, Svensson, Vallelonga, Vinther, Walker, Wheatley and Wistrup2014): stadial periods marked as GS, interstadial periods as GI; GI periods are additionally highlighted as vertical bands.
We aim to analyze results within the 15–50 cal kBP range. However, the database also includes younger age determinations to allow for tracking trends in more recent periods. Thus, Figure 2 also includes portions of probability distributions for dates younger than 15 cal kBP. For the Dinkel River valley (eastern Netherlands), the youngest date in our database is GrN-13403, 11,285±25 BP.
The shape of the PDF curves for the other regions suggests a sharp increase in the number of dates with the onset of warming during GI-1. Furthermore, for all three regions, the cold event GS-1 is clearly marked by the PDF minima, while adjacent distribution maxima accurately reflect the beginning and end of GS-1.
In each distribution, it is evident that PDF maxima correlate with interstadials (GI).
For the Dinkel River valley, we have a relatively large number of the oldest dates, and the resulting PDF distribution, with a high and broad maximum (47–38 cal kBP), appears significantly different from the other regions. However, it is worth examining which date clusters contribute to the summed PDF curves.
A detailed analysis of the obtained curves is presented in the Discussion section.
Analysis of PDF Curves for Individual Regions
For each region, in addition to the summed probability density function (PDF) of all dates (part B), we have presented the calibration results as 68.3% confidence intervals in the form of horizontal bars (part A) and examined the clustering of dates against INTIMATE stratigraphy (part E, with vertical bands representing interstadials). These results distinguished subgroups were, and their PDF distributions were generated (part C). Additionally, a weighted mean was calculated for each subgroup, and the calibration results of these means are presented in the graphs (part D)—see Figures 3–6.
Figure 3. Analysis of the radiocarbon dataset for the Eastern Netherlands (Dinkel River valley). A – 68.3% confidence intervals of calibrated individual dates. B – Summed probability density function (PDF). C – PDF distributions for date subgroups (the number of dates in each subgroup is indicated above each peak). D – Calibration results of weighted means for each subgroup. E – Same as in Figure 2.
Figure 4. Analysis of the radiocarbon dataset for the Eastern Germany. A–E – same as in Figure 3.
Figure 5. Analysis of the radiocarbon dataset for S Poland, Subcarpathian Basins. A–E – same as in Figure 3.
Figure 6. Analysis of the radiocarbon dataset for S Poland, Carpathians. A–E – same as in Figure 3.
When interpreting the shape of the PDF graphs, it is essential to remember that sharp and narrow peaks may result from individual dates with low uncertainty (mainly AMS dates) as well as from steep sections of the calibration curve (see Michczyński and Michczyńska Reference Michczyński and Michczyńska2006). Therefore, it is not the height of the peak but its presence that indicates changes recorded in the fluvial environment. Conversely, declines in the PDF curves may result either from periods during which no significant fluvial changes occurred (e.g., dry and cold periods) or from an insufficient number of dated samples.
Eastern Netherlands
For the Dinkel River valley, 87 radiocarbon dates are available. However, most of these dates are older than 30 cal kBP, with only six dates younger than 16 cal kBP. Therefore, a detailed analysis was conducted for the 50–30 cal kBP period, as presented in Figure 3.
One date (GrN-1217, 36,300+3,200/-2,300 BP) was excluded from the figure because its uncertainty was significantly larger than other dates of similar age.
Van Huissteden (Reference Van Huissteden1990), from whose work the database was derived, noted that some samples might be contaminated, as the radiocarbon dates obtained from NaOH-soluble and residue fractions differed. To investigate this, Figure 3B includes a PDF curve with contaminated samples removed. The results show that excluding these dates does not significantly alter the overall PDF shape, so the analysis was performed on the full dataset.
Examining Figure 3A, it is evident that the length of the 68.3% confidence intervals increases with age. This is a statistical regularity, as radiocarbon dating uncertainties increase with age. This effect is noticeable across all datasets analyzed from the three countries.
-
• Based on the shape of the PDF curve (part B), date clustering (part A), and INTIMATE stratigraphy (part E), 17 subgroups were distinguished within the period analyzed
The vast majority of date subgroups are associated with GS/GI or GI/GS transitions. For long interstadial periods (GI-14–GI-13, GI-12, GI-8), distinct subgroups are observed both at the beginning and end of the interstadials. Only two subgroups are associated with GS-12 and GS-11, and two individual dates correspond to GS-7 and GS-5.2.
The calibrated distributions of weighted means for each subgroup allow for determining the timing of environmental changes recorded in the sediments of the Dinkel River valley with lower uncertainty than individual dates.
Eastern Germany
For eastern Germany, the database consists of 64 radiocarbon dates, with the youngest date being 10,100±160 BP and the oldest 46,700±1,600 BP.
The results of the analysis for eastern Germany are presented in Figure 4. This figure does not include five dates with significantly larger uncertainties than neighboring dates: 33,105±5,000 BP, 32,320±2,260 BP, 29,550±1,560 BP, 29,380±2,400 BP, and 26,110±4,000 BP. Due to their high uncertainties, these dates would only broaden the probability distributions for the respective subgroups.
Following the same approach as with the Netherlands dataset, 24 subgroups were distinguished.
Similar to the Netherlands dataset, the eastern German dataset reveals a tendency for 14C dates to cluster around GS/GI and GI/GS transitions.
Southern Poland, Subcarpathian Basins
Out of the 257 collected radiocarbon age determinations, only about 80 dates are older than 15 cal kBP after calibration. The results are presented in Figure 5. The analysis was conducted similarly to the dataset from the Dinkel River valley (Netherlands). However, dates with large uncertainties were not excluded to maintain the available sample size. As a result, the distributions for subgroups are broader than in the Netherlands and Germany datasets.
A clear increase in the number of 14C dates is visible with the warming corresponding to GI-1 and the Holocene. The fewest dates occur within 25–20 cal kBP, corresponding to the maximum extent of the Vistulian ice sheet (GS-3 – GS-2c) and GS-9.
Following the same approach as with the Netherlands dataset, 20 radiocarbon date subgroups were distinguished. Their PDF distributions are presented in part C.
Southern Poland, Carpathians
The database consists of 64 dates, but only 38 dates are older than 15 cal kBP after calibration. The results are presented in Figure 6. As with the Subcarpathian Basins dataset, dates with large uncertainties were not excluded to maintain the available sample size, leading to broader distributions for subgroups compared to the Netherlands and Germany datasets.
Following the same approach as with the Netherlands dataset, 17 subgroups were distinguished.
For this dataset, the importance of precise dates is particularly evident (see the group correlated with GI-3).
Discussion
Similarities and differences between the studied regions
Similarities
-
1. In all study areas, there is a clear tendency for radiocarbon dates to cluster around GS/GI and GI/GS transitions.
-
2. All regions show a decrease in the number of dates during GS-3 and GS-2.
-
3. The number of dates associated with stadials (GS) is generally low.
-
4. In eastern Germany and the two regions in southern Poland, there is a sharp increase in the number of dates from the GS-2a/GI-1 transition to the Holocene, with a distinct minimum during GS-1.
-
5. The greatest similarity is observed between the summed and subgroup distributions of eastern Germany and the Subcarpathian Basins.
Differences
-
1. Only in the Netherlands dataset are dates correlated with the short interstadial GI-9. This may be because the Dutch dates have the smallest uncertainties.
-
2. The longest gap in radiocarbon dates (30–15 cal kBP) occurs in the Netherlands dataset, likely because the data come from a single small river valley.
Possible causes of differences in PDF curve shapes across regions:
-
1. Geographical location, i.e., distance from oceanic influences, increasing continentality with longitude, and the distance of studied sites from the ice-sheet margin.
-
2. Number of river valleys and radiocarbon dates included in the analysis.
-
3. Size of the rivers and the scale of fluvial processes occurring within them.
-
4. Dated material and chemical fraction. The most commonly dated materials were peat and organic mud/silt. The collected dates (with the exception of a few modern ones) were obtained using conventional radiocarbon dating techniques, and the entire material was dated without fraction separation by sieving.
-
5. The impact of contamination. Unfortunately, only for some samples were both the residue and extract fractions analyzed (mainly for samples from the Netherlands and partly from Germany and in one case for Poland), allowing for inferences about possible contamination. The analyses utilized dates obtained from the residue fraction (see supplementary materials).
-
6. Accuracy of the dates used in the analyses. For example, based on the results presented, it is difficult to determine whether there is a time shift in the onset of organic sediment accumulation linked to GI events, depending on geographic location. More precise 14C analysis would be needed to examine this, but most of the database consists of conventional radiocarbon dates.
-
7. There are differences between laboratories that performed the age determinations. Inter-laboratory errors are generally small and only become significant in high-precision reconstructions using AMS dates. Ideally, all dates should be obtained from a single laboratory.
Summary graphs
Figure 7 presents the results of the calibration of means for individual date subgroups within the analyzed regions. The calibration results are shown as 68.3% confidence intervals (horizontal bars). The highest dating precision is clearly observed for the Netherlands. Therefore, in this region, the clustering of dates at the GS/GI and GI/GS transitions is most clearly visible (See also supplementary materials, last sheet of the Excel file).
Figure 7. Results of the calibration of means for individual date subgroups within the analyzed regions presented as 68.3% confidence intervals (horizontal bars).
Since common trends are visible across all analyzed regions, a summed PDF graph was created (Figure 8). The graph presents calibrated individual dates and PDF curves for:
-
• All 503 radiocarbon dates from the database.
-
• A subset of 444 dates, excluding those with high uncertainties (i.e., those exceeding: 2% of 14C age values for dates <15 kBP, 5% of 14C age values for dates <25 kBP, 7% of 14C age values for dates <40 and 10% for older dates).
Figure 8. Summed PDF graph for all analyzed regions. The upper graph includes all 503 collected radiocarbon age determinations. The lower graph presents results for 444 dates, excluding those with significantly higher uncertainties than similar aged dates.
The exclusion of high-uncertainty dates affects only the relative height of the PDF curve but does not alter its shape. This makes it easier to analyze date clusters based on the calibration results of individual dates against the INTIMATE stratigraphy framework.
The same trends observed in individual regions are also present in the aggregate distribution, with common patterns becoming more pronounced. The PDF maxima in the aggregate distribution correlate with interstadials (GI) or GI/GS transitions. Interstadials younger than 40 cal kBP, as well as GI-11 and the end of GI-12, are the most clearly marked. Notably, the GS-2c/GS-2b and GS-2b/GS-2a transitions are also reflected as peaks in the PDF curve.
Paleoenvironmental perspective
General Observations can be listed as follows:
-
− Interstadials (GI) are periods with more favorable climatic conditions (warmer and wetter), which can potentially promote the formation of organic deposits. However, mineral sediments (e.g., channel deposits) were also formed during GI.
-
− Stadials (GS) were periods of colder and drier conditions, that were less favorable—or even unfavorable for organic sediment deposition. Instead, they were dominated by mineral accumulation. Periglacial structures also marked more severe cold conditions during GS.
-
− Fluvial erosion phases recorded in sandy deposits may also have paleoclimatic significance. These processes were likely more intense at the beginning and end of GS periods due to:
-
○ Initial GS phases: Vegetation cover became sparse, often with open communities; Erosion associated with braided river systems characterized by a positive sediment budget.
-
○ Final GS phases: Vegetation cover became denser, which could influence water availability in river channels; Channel erosion took place during the transition from braided to meandering river types.
-
-
− Fluvial sediment profiles can provide additional paleoenvironmental information. Sandy sequences of channel deposits often show erosional surfaces or frost-related features.
-
− Middle Pleniglacial sequences comprise often organic deposits of palaeochannel fill or thermokarst lakes. Indicating cool and wetter climatic conditions of tundra rivers.
-
− Upper Pleniglacial sediment sequences mainly comprise mineral (sandy) deposit, indicating cold and dry palaeoenvironmental conditions.
-
− In foreland areas and intermontane basins, aggradation generally prevails, resulting in the formation of alluvial fill terraces. In contrast, mountainous regions are characterized by significant sediment supply and a predominance of erosional processes, leading to the development of strath terraces (erosional-accumulation terraces).
-
− In the western part of the Carpathians, the influence of moist oceanic air masses is evident, contributing to higher precipitation levels. East of the Dunajec River catchment, the climate becomes increasingly continental, which affects the hydrological regime of rivers, the types and dynamics of flood events, and, more broadly, fluvial sedimentation patterns.
-
− The morphology of river channels also differs considerably across these regions, including channel gradient and the grain-size composition of transported material—factors that are closely linked to the underlying geological structure.
Interpleniglacial climate variability
Summary of basic literature data:
-
− The Interpleniglacial period had a cool but slightly wetter climate compared to the Lower and Upper Pleniglacial (Ran Reference Ran1990; Sarala et al. Reference Sarala, Väliranta, Eskola and Vaikutiené2016; Britzius et al. Reference Britzius, Dreher, Maisel and Sirocko2024).
-
− This ∼30,000-year period consisted of many warming and cooling cycles (Dansgaard et al. Reference Dansgaard, Johsen, Clausen, Dahl-Jensen, Gundestrup, Hammer, Hvidberg, Steffensen, Sveinbjornsdottir, Jouzel and Bond1993; Rasmussen et al. Reference Rasmussen, Bigler, Blockley, Blunier, Buchardt, Clausen, Cvijanovic, Dahl-Jensen, Johnsen, Fischer, Gkinis, Guillevic, Hoek, Lowe, Pedro, Popp, Seierstad, Steffensen, Svensson, Vallelonga, Vinther, Walker, Wheatley and Wistrup2014).
-
− These fluctuations are recorded in only a few stratigraphic pollen sequences from Western Europe (Britzius et al. Reference Britzius, Dreher, Maisel and Sirocko2024; Guiot et al. Reference Guiot, Pons, de Beaulieu and Reille1989; Grootes and Stuiver Reference Grootes and Stuiver1997).
-
− Classic profiles from the eastern Netherlands and Germany (Behre and van der Plicht Reference Behre and Van der Plicht1992; Van der Hammen and Wijmstra Reference Van der Hammen and Wijmstra1971) demonstrate this climate variability through alternating deposits:
-
○ Interstadials (warm periods) are represented by organic deposits.
-
○ Stadials (colder climate oscillations) are represented by mineral (sandy) deposits.
-
-
− However, fluvial sediments from many Pleniglacial rivers do not necessarily reflect direct climate fluctuations. Instead, they could represent:
-
○ Facies changes in deposits (overbank organic sedimentation in palaeochannel fill or thermokarst lakes, channel sandy deposits or overbank sandy-silty natural levees deposits).
-
○ Shifts in river sedimentation styles, from braided to meandering/anastomosing rivers (Mol Reference Mol1997; Van Huissteden Reference Van Huissteden1990).
-
-
− Palynological data from continuous profiles, such as Eifel, Germany (Britzius et al. Reference Britzius, Dreher, Maisel and Sirocko2024) show:
-
○ Increased presence of boreal forest taxa during GI.
-
○ Higher abundance of grassland-steppe taxa during GS.
-
○ Even short-term GI events are reflected in vegetation changes.
-
-
− Palynological analysis of organic deposits within fluvial sequences provides limited climatic information:
-
○ These sediments formed under similar climate conditions, generally supporting open forest, forest-steppe and steppe-tundra ecosystems
-
○ Even if two organic horizons are stratigraphically superimposed, their correlation with specific GI periods remains uncertain without precise radiocarbon dating.
-
Conclusions
The study presented in this article on fluvial activity during the Interpleniglacial and Upper Pleniglacial of the Weichselian Cold Stage demonstrates that rhythmic climate fluctuations played a crucial role in the evolution of river valleys in the extraglacial zone. These fluctuations led to alternating erosional and aggradational phases in river valleys. Valley deepening was driven not only by changes in the erosional base but also by increasing aridification, particularly in the Upper Pleniglacial.
Thicker organic deposits are rare in alluvium from this period, making it difficult to establish a precise chronology of fluvial events. Nevertheless, radiocarbon dating and the analysis of grouped dates provide valuable insights. The findings from the analyzed regions can be summarized as follows:
-
1. Eastern Netherlands: Interstadials from GI-14 to GI-6 are clearly represented, mainly through GS/GI and GI/GS transitions.
-
2. Eastern Germany: GS/GI and GI/GS transitions associated with interstadials from GI-12 to GI-3 are well represented, except for the short interstadial GI-9. A period with few radiocarbon dates is observed between 23.0–15.0 cal kBP.
-
3. Subcarpathian Basins: Similar to eastern Germany, GS/GI and GI/GS transitions related to interstadials from GI-12 to GI-3 are recorded, except for GI-9. A clear gap in radiocarbon dates is observed between 25.0–22.5 cal kBP, which is shorter than the entire Upper Pleniglacial (MIS2).
-
4. Carpathians: This region has the fewest dates in the analyzed time range. However, interstadials from GI-12 to GI-7 are well represented, except for GI-9, GI-4, and GI-3. A longer period without radiocarbon dates (23.0–15.0 cal kBP) is observed. The higher hypsometric position of Carpathian rivers compared to the basins suggests that they may have been more influenced by colder or drier air masses, which could have negatively impacted the formation of organic sediments.
-
5. The recurring clustering of dates at GS/GI and GI/GS transitions across different regions of Europe suggests that these transitions reflect supra-regional paleoclimatic conditions that influenced river system dynamics.
-
6. The study confirms that the number of warming-cooling cycles recorded in sediments aligns with Greenland ice-core data. This supports the hypothesis that fluvial records can serve as proxies for past climate variability.
The study provides new insights into how climate fluctuations influenced fluvial processes in different regions. Future studies should incorporate a greater number of AMS radiocarbon dates to improve chronological resolution. High-precision dating would help refine correlations between regional fluvial responses and global climate events.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/RDC.2025.10162
Acknowledgments
The authors are grateful to the Associate Editor Pieter Grootes and reviewers—Associate Editor Irka Hajdas and the anonymous reviewer—for their insightful comments and suggestions, which greatly enhanced the quality of the manuscript.
Participation of two authors (DJM and AM) in the 4th International Radiocarbon in the Environment Conference has been supported by EU funds FSD–10.25 Development of higher education focused on the needs of the green economy European Funds for Silesia 2021–2027: The modern methods of the monitoring of the level and isotopic composition of atmospheric CO2 (project no.FESL.10.25-IZ.01-06C9/23-00) implemented at the Silesian University of Technology (2024–2026). This work for one of author (DW) was supported by the internal Polish Geological Institute – National Research Institute, research grant project no. 62.9012.2520.00.0.
Open access
