Synchronizing the Western Gotland Basin (Baltic Sea) and Lake Kälksjön (central Sweden) sediment records using common cosmogenic radionuclide production variations

Abstract

Multi-archive studies of climate events and archive-specific response times require synchronous time scales. Aligning common variations in the cosmogenic radionuclide production rate via curve fitting methods provides a tool for the continuous synchronization of natural environmental archives down to decadal precision. Based on this approach, we synchronize ¹⁰Be records from Western Gotland Basin (WGB, Baltic Sea) and Lake Kälksjön (KKJ, central Sweden) sediments to the ¹⁴C production time series from the IntCal20 calibration curve during the Mid-Holocene period ~6400 to 5200 a BP. Before the synchronization, we assess and reduce non-production variability in the ¹⁰Be records by using ¹⁰Be/⁹Be ratios and removing common variability with the TOC record from KKJ sediments based on regression analysis. The synchronizations to the IntCal20 ¹⁴C production time scale suggest decadal to multi-decadal refinements of the WGB and KKJ chronologies. These refinements reduce the previously centennial chronological uncertainties of both archives to about ± 20 (WGB) and ±40 (KKJ) years. Combining proxy time series from the synchronized archives enables us to interpret a period of ventilation in the deep central Baltic Sea basins from ~6250 to 6000 a BP as possibly caused by inter-annual cooling reducing vertical water temperature gradients allowing deep water formation during exceptionally cold winters.

Keywords

chronology Mid-Holocene climate sediment archives time scale synchronization western Baltic region

Introduction

Paleoclimate reconstructions provide insights into past climate dynamics and, thereby, improve our understanding of the climate system (Rasmussen et al., 2014; von Grafenstein et al., 1999). Multi-archive investigations of these reconstructions allow us to study leads and lags in proxy responses to climate changes providing information about their temporal progression and spatial gradients, archive-specific thresholds and the driving mechanisms (Czymzik et al., 2020; Fleitmann et al., 2009). However, due to the often considerable uncertainties of the archives’ individual chronologies, such detailed studies require independent synchronization tools.

For sedimentary archives, volcanic ash (tephra) deposits provide such a tool (Blockley et al., 2014; Wulf et al., 2013). The geochemical fingerprint of tephra deposits provides stratigraphic isochrones enabling the alignment of different sediment archives (Lane et al., 2015). However, this method is limited by the stochastic occurrence of large volcanic eruptions and regionally restricted fallout patterns. Matching globally common production rate variations of cosmogenic radionuclides using curve fitting methods can provide a continuous synchronization tool for environmental archives (Czymzik et al., 2016; Mellström et al., 2015; Muscheler et al., 2014; Snowball et al., 2010).

Cosmogenic radionuclides like ¹⁰Be and ¹⁴C are produced in Earth’s upper atmosphere by the interaction of atmospheric target nuclei with incident galactic cosmic rays (Lal and Peters, 1967). Varying helio- and geomagnetic shielding of these galactic cosmic rays result in a characteristic cosmogenic radionuclide production and fallout pattern imprinted in environmental archives around the globe (Aldahan and Possnert, 1998; Lal and Peters, 1967; Simon et al., 2020). Placing sedimentary ¹⁰Be records on the precise ¹⁴C production rate time scale from the IntCal20 calibration curve using curve fitting methods enables us to synchronize these records and reduce the uncertainties of the connected chronologies (Czymzik et al., 2018, 2020).

A challenge of this method is the unequivocal detection of the ¹⁰Be production signal due to postproduction influences mainly introduced by the redeposition of previously deposited “old” ¹⁰Be, varying atmospheric circulation and sedimentology (Balco et al., 2021; Czymzik et al., 2015; Simon et al., 2016). One way to reduce these influences is the analysis of ¹⁰Be/⁹Be-ratios (Wittmann et al., 2012). Introduced into a natural environment ¹⁰Be and ⁹Be behave chemically identical. However, while ¹⁰Be is produced in the atmosphere, ⁹Be originates from weathering crustal rocks (Simon et al., 2020; Wittmann et al., 2012). Therefore, common variations of both isotopes in a sediment archive can be assumed to mainly reflect transport of previously deposited “old” ¹⁰Be and ⁹Be within the lake/catchment setting and the ¹⁰Be/⁹Be-ratio an environment-corrected record of the atmospheric ¹⁰Be production rate (Savranskaia et al., 2021). Another correction method comprises regression analyses between ¹⁰Be records and environmental proxy time series from the same sediment core (Czymzik et al., 2018). Thereby, it is assumed that common variability with a proxy record represents an environmental bias in the ¹⁰Be record. However, due to the sole statistical nature, the results of this approach require a discussion about the underlying mechanisms.

Previous studies on the Lake Kälksjön (KKJ, central Sweden) sediment sequence successfully applied ¹⁴C wiggle-match dating to connect its chronology to the IntCal time scale around ~8200 and between 3000 and 2000 a BP, with errors of down to ±20 years (Mellström et al., 2015; Snowball et al., 2010). In the following, we attempt to synchronize ¹⁰Be records from brackish Western Gotland Basin (WGB, Baltic Sea) and terrestrial KKJ sediments to the ¹⁴C production time-series inferred from the IntCal20 calibration curve during the Mid-Holocene period ~5200 to 6400 a BP. Favorable for the synchronization, this period includes a succession of distinct centennial cosmogenic radionuclide production peaks. Considering the high replication of the underlying ¹⁴C data and dendrochronological precision of this part of the calibration curve, we use it as an absolute dated target time series (Reimer et al., 2020). Before the synchronization, the ¹⁰Be data from WGB and KKJ sediments were corrected for the average atmospheric residence time of ~1 year (Raisbeck et al., 1981) and possible non-production influences. Based on the synchronized records, we investigate the likely driving mechanisms of a period of ventilation and deposition of well-mixed sediments in the central Baltic Sea basins ~6250 to 6000 a BP (Moros et al., 2020; Warden et al., 2017).

Study sites

The presently brackish Baltic Sea formed ~8000 years ago mainly as a result of post-glacial sea-level rise (Figure 1) (Björck, 1995). Competing inflows of saline waters from the North Sea via the shallow Danish Straits and freshwater from rivers and precipitation led to the establishment of a strong pycnocline (Emeis et al., 2003). Its morphologically complex basin is subdivided through sills and connected by channels (Leppäranta and Myrberg, 2009). Postglacial isostatic land uplift successively modifies water depths and shoreline positions (Berglund et al., 2005; Ekman, 1996). Continuous sediment accumulation is restricted to the deeper areas of the Baltic Sea, while wave-induced sediment transport or erosion becomes dominant in shallower and coastal regions (Winterhalter, 1992). Only within the deep basins finely laminated organic-rich sediments can deposit under predominantly oxygen-deficient and sulfidic bottom water conditions (Andrén et al., 2000; Zillén et al., 2008). An exception is the modern period where the effects of anthropogenic forcing and global warming favored anoxia and the deposition of fine laminated sediments also in some shallow coastal basins (Jokinen et al., 2018). The Baltic Sea has a surface area of 377,000 km² and catchment size of 1,641,650 km². Its mean depth is 54 m and the deepest point is located in the funnel-shaped Landsort Deep (459 m) (BACC, 2015). The WGB is located in the central Baltic Sea, off south-eastern Sweden (Figure 1).

Figure 1.

Geographical locations of sediment cores M86-1a/33-4GC from the Western Gotland Basin (WGB) and P435-2-1 from the Eastern Gotland Basin, as well as Lake Kälksjön (KKJ) in central Sweden. Map modified after Kaiser et al. (2017).

KKJ is situated in central Sweden (60°09′N/13°02′E; 97 m a.s.l.), ~250 km west of the Baltic Sea coast (Figure 1). KKJ was isolated from ancient Lake Vänern ~9600 years ago through isostatic uplift (Czymzik et al., 2023; Mellström et al., 2015; Stanton et al., 2010). The present basin has a catchment size of 4 km², surface area of 0.42 km² and maximum water depth of 13.6 m. Four minor creeks discharge into KKJ from the forested northern and eastern catchment (Stanton et al., 2010). The lake’s only outflow in the West was artificially incised in 1878 CE (Czymzik et al., 2023).

Material and methods

Sediment cores and initial chronologies

Sediment core M86-1a/33-4GC (58°21.9N/17°50.0E) was retrieved from the WGB at 101 m water depth using a gravity corer during a cruise with RV Meteor in 2011 (Figure 1). Age constraints for the investigated sediment interval from 189 to 219 cm core depth were transferred from the radiometrically dated Eastern Gotland Basin sediment core P435-2-1, through matching common variability in the loss-on-ignition (LOI) and XRF Br organic matter proxies (Warden et al., 2017).

Sediment core KKJ19 was retrieved from the central basin of the lake in 2019 using an UWITEC piston corer (Czymzik et al., 2023). The age-depth model for sediment core KKJ19 was built using 8 calibrated AMS ¹⁴C ages, ¹³⁷Cs and ²⁴¹Am measurements, as well as lithological marker layers of known age (Czymzik et al., 2023).

Beryllium extraction and measurements

For the extraction and measurement of ¹⁰Be and ⁹Be records at ~20 to 30-year resolution, series of samples were collected at 0.5 cm step-size from sediment core M86-1a/33-4GC and 1 cm resolution from sediment core KKJ19. After spiking with 0.2 mg ⁹Be carrier, Be was leached from 0.2 g freeze-dried and homogenized sediment samples overnight with 32 vol% HCl and H₂O₂ (Henken-Mellies et al., 1990). The resulting solution was separated from the remaining material through centrifuging. Then, hydroxides were precipitated at pH 10 using 25 vol% NH₄OH and separated from the solution by centrifuging. Be was co-precipitated with Fe using 25 vol% NH₄OH and then re-dissolved by increasing the pH to 14 with NaOH. Be(OH)₂ was precipitated from the remaining solution using 25 vol% NH₄OH at pH 8–10 and washed two times with ultra-pure H₂O. After oxidizing at 600°C in a muffle furnace, BeO was mixed with Nb and pressed into cathodes for AMS measurements. Final ¹⁰Be concentrations were calculated from measured ¹⁰Be/⁹Be ratios, normalized to the S2007N (¹⁰Be/⁹Be = 28.1 × 10⁻¹²) and S2010N (¹⁰Be/⁹Be = 3.3 × 10⁻¹²) reference standards (Christl et al., 2013).

⁹Be was extracted from aliquots of the same sediment samples as for ¹⁰Be. The procedure includes treatment with 15 vol% HCl for 1 h on a shaker, centrifuging, decanting and cleaning through a 0.4 µm mesh-size filter. Measurements of ⁹Be contents were conducted with a quadrupole ICP-MS (iCAP Q, Thermo Fisher Scientific) using matrix-matched external calibration. Rhodium served as internal standard to compensate for instrument fluctuations and remaining matrix effects. Precision and accuracy were checked by the international sediment reference materials SGR-1b (USGS) and TH-2 (Canada) and were better than 1.6% and 3.9%, respectively. Because no certified Be contents were available for the here-applied HCl extraction, both materials were subjected to total acid digestion using a HNO₃–HF–HClO₄-mixture, to determine the analytical error (Dellwig et al., 2019).

XRF element profiles and loss on ignition

X-ray fluorescence (XRF) measurements at 200 µm resolution were conducted on WGB sediment core M86-1a/33-4GC using an ITRAX core scanner with a Cr tube and an energy-dispersive SDD detector (Croudace et al., 2019). The 10 mm wide X-ray beam was orientated parallel to the laminations, to minimize the effects of sediment micro-disturbances. Individual measurements were performed at 30 kV and 30 µA, with a dwell time of 15 s. Loss on ignition (LOI) was determined at 1 cm resolution from freeze-dried samples of sediment core M86-1a/33-4GC, combusted for 3 h at 550°C. Sediment core M86-1a/33-4GC includes no considerable amounts of diagenetic Mn carbonate that would bias organic matter estimates from LOI data (Häusler et al., 2017).

Correcting for environmental influences on ¹⁰Be concentrations

To detect and correct for non-production-related biases in the ¹⁰Be records from WGB and KKJ sediment cores, two approaches were applied. First, we use ratios of the isotopes ¹⁰Be (meteoric production and redeposition) and ⁹Be (released from weathering rocks and redeposition) to remove environmental biases from our ¹⁰Be concentration records (Simon et al., 2020; Wittmann et al., 2012). Second, we compare the ¹⁰Be concentration records with proxy time series from the same archives, reflecting the main sediment constituents (organic matter: WGB = LOI, KKJ = TOC (Czymzik et al., 2023); detrital material: WGB = Ti/Br, KKJ = Ti (Czymzik et al., 2023). Ti/Br can be used to reduce the effects of varying organic matter contents on XRF Ti counts in marine sediments (Ziegler et al., 2008). In case of a significant correlation, we removed the covariance from the ¹⁰Be concentration record using linear regression (only TOC in KKJ = ¹⁰Be_TOC) (Adolphi and Muscheler, 2016; Czymzik et al., 2020).

¹⁴C production rates

¹⁴C production rates used in this study were calculated using a box-diffusion carbon cycle model assuming no changes in the global carbon cycle over the investigated period (Muscheler et al., 2005; Siegenthaler, 1983). The uncertainties of the IntCal20 calibration curve were included by using 100 posterior realizations of possible atmospheric ¹⁴C curves, obtained by fitting Bayesian splines to the ¹⁴C data underlying IntCal20 (Heaton et al., 2020; Reimer et al., 2020). The reported ¹⁴C production rate reflects the average obtained from the 100 realizations.

Time scale synchronization and propagating uncertainties

The ¹⁰Be/⁹Be-ratio time series from both archives, as well as the TOC-corrected ¹⁰Be record from KKJ sediments were semi-automatically synchronized to the atmospheric ¹⁴C production record from the IntCal20 calibration curve using the MATCH software working with the global optimal fit (Lisiecki and Lisiecki, 2002; Reimer et al., 2020). MATCH divides the records into small segments and calculates all possible alignments to find the optimal one. The alignment score is expressed as the square of the difference between the two signals (Lisiecki and Lisiecki, 2002). Before the synchronization, all data were detrended and z-score transformed by subtracting the mean and dividing by the standard deviation (Lisiecki and Lisiecki, 2002). Up to ±25% changes in sedimentation rates were allowed to account for the effects of chronological and sampling uncertainties and all realizations with at least 1-year difference included in the calculation of the final mean synchronization results. Error margins for the synchronization procedure were determined by applying root mean squared error (RMSE) estimates, considering two identified uncertainties and using the following formula (Grant et al., 2012):

R M S E = \sqrt (a^{2} + b^{2})

(a) Synchronization uncertainty, defined as the maximum difference between the individual synchronization realizations.

(b) Sample resolution of the ¹⁰Be records.

Results

Beryllium inventories and sediment composition

Concentrations of ¹⁰Be and ⁹Be were measured in 54 samples from WGB sediments covering the interval from 189 to 219 cm core depth. ¹⁰Be concentrations vary between 179.4 and 305.3 × 10⁶ a g⁻¹, around a mean of 248.7 × 10⁶ a g⁻¹ (Figure 2). The ¹⁰Be record from WGB sediments shows cm-scale excursions and amplitude changes of ~50% throughout the record (Figure 2). One ¹⁰Be data point from an earlier series was excluded from the analyses, due to uncertainties regarding the added ⁹Be tracer. Correlations of ¹⁰Be in WGB sediments with the organic (LOI: r = 0.04, p = 0.41) and detrital matter (Ti/Br: r = 0.15, p = 0.12) proxies are very low (Figure 3). ⁹Be contents in WGB sediments range from 2.6 to 4 × 10¹⁶ a g⁻¹ (mean 3 × 10¹⁶ a g⁻¹) and show a slightly increasing trend between 189 and 217 cm core depth (Figure 2).

Figure 2.

Investigated parts of Western Gotland Basin (WGB) and Lake Kälksjön (KKJ) sediment cores with Be and selected proxy data. (a) Photographs of the investigated parts of sediment cores M86-1a/33-4GC (intervals with laminated and homogeneous sediments are indicated) and KKJ19, (b) ¹⁰Be, (c) ⁹Be, (d) ¹⁰Be/⁹Be, (e) organic matter proxies (WGB: LOI, KKJ: TOC), (f) detrital matter proxies (WGB: Ti/Br, KKJ: Ti).

Figure 3.

Correlations of ¹⁰Be records from Western Gotland Basin (WGB) and Lake Kälksjön (KKJ) sediments with proxy records of changing organic and detrital matter contents. (a) Organic matter proxy loss on ignition (LOI) and ¹⁰Be in WGB sediments. (b) Detrital matter proxy Ti/Br and ¹⁰Be in WGB sediments. (c) Organic matter proxy total organic carbon (TOC) and ¹⁰Be in KKJ sediments. (d) Detrital matter proxy Ti and ¹⁰Be in KKJ sediments.

¹⁰Be and ⁹Be concentrations were measured in 33 samples from KKJ sediments. ¹⁰Be concentrations range from 547.1 to 968.8 × 10⁶ a g⁻¹, around a mean of 690.7 × 10⁶ a g⁻¹ (Figure 2). The ¹⁰Be record reveals smaller variations between 290 and 350 cm core depth and a larger excursion from 280 to 290 cm core depth (Figure 2). ¹⁰Be concentrations from KKJ sediments reveal a significant correlation with the TOC record from this archive (r = 0.58, p < 0.01) (Figure 3). The correlation between ¹⁰Be and the detrital matter proxy Ti in KKJ sediments is very low (r = 0.02, p = 0.46) (Figure 3). ⁹Be concentrations in KKJ sediments reach from 0.87 to 1.59 × 10¹⁷ a g⁻¹ (mean 1.1 × 10¹⁷ a g⁻¹) (Figure 2). Blank corrections were negligible for the ¹⁰Be data from WGB and KKJ sediments.

Discussion

Mechanisms of ¹⁰Be deposition in WGB and KKJ sediments

Changing sediment composition and the redeposition of “old” ¹⁰Be can bias the expected ¹⁰Be production signal in sediment records (Berggren et al., 2010; Czymzik et al., 2018). This chapter aims at disentangling the mechanisms of ¹⁰Be deposition in WGB and KKJ sediments.

Very low correlations with ⁹Be, the organic and detrital matter proxies, as well as resemblances with the ¹⁰Be/⁹Be and ¹⁴C production records indicate minor environmental influences and a good preservation of the ¹⁰Be production rate variations in WGB sediments (Figures 2 –5). Amplitude changes of ~60% for most of the record match the expected production rate variations during centennial-scale solar variability changes (Muscheler and Heikkilä, 2011) (Figure 4).

Figure 4.

Correcting for environmental influences on ¹⁰Be deposition in (a) Western Gotland Basin (WGB) and (b) Lake Kälksjön (KKJ) sediments using ¹⁰Be/⁹Be-ratios and regression analysis with the significantly correlated TOC record from KKJ sediments.

Figure 5.

Synchronization of ¹⁰Be/⁹Be record from Western Gotland Basin (WGB) sediments with ¹⁴C production rates from the IntCal20 calibration curve (Reimer et al., 2020). (a) ¹⁰Be/⁹Be record (blue) synchronized to ¹⁴C production rates (red). The dashed line depicts the ¹⁰Be/⁹Be record on its original time scale. The significance of the correlation coefficient takes into account the effects of autocorrelation in the time series (Ebisuzaki, 1997). (b) Chronological changes applied to the ¹⁰Be/⁹Be record during the synchronization with root mean square error (RMSE) estimates (Grant et al., 2012). (c) Moving correlation between the synchronized ¹⁰Be/⁹Be and ¹⁴C production records.

Two exceptions are the short sediment intervals around 192 cm and from 210 to 214 cm core depth (Figure 2). Between 210 and 214 cm core depth (~6300 to 6100 a BP), no decadal-scale variability is preserved in the ¹⁰Be and ¹⁰Be/⁹Be records from WGB sediments, in contrast to the ¹⁴C production time-series (Figures 2 and 5). This interval corresponds broadly to a non-laminated sediment unit in the WGB sediment core that is a common stratigraphic feature in the deep central Baltic Sea basins (Warden et al., 2017; Zillén et al., 2008). It is interpreted to be mainly triggered by water turbulences during ventilation causing sediment redeposition and mixing (Moros et al., 2020; Warden et al., 2017). Assumedly, sediment mixing also led to a smoothing the ¹⁰Be and ¹⁰Be/⁹Be records from WGB sediments during this interval.

In turn, a minimum in the ¹⁴C production record without a corresponding response in the WGB ¹⁰Be record around 192 cm core depth (~5300 a BP) can be corrected for using ¹⁰Be/⁹Be-ratios (Figures 4 and 5). Therefore, the main reason likely is an enhanced contribution of “old” ¹⁰Be to the coring site that is paralleled by a similar enrichment of ⁹Be, without substantial sediment mixing at the coring site (Figure 5).

In contrast to ¹⁰Be in WGB sediments (except for 192 cm core depth), ¹⁰Be concentrations in KKJ sediments reveal significant correlations (p < 0.01) with the ⁹Be and TOC records from this archive (Figure 3). These correlations point to a strong environmental influence on ¹⁰Be deposition in the lake. TOC contents in KKJ sediments during the last ~9600 years were interpreted to predominantly reflect productivity changes controlled by North Atlantic Oscillation (NAO) driven winter temperature influences on ice cover duration and growth season length (for a detailed discussion see Czymzik et al., 2023). Sedimentary TOC contents are enhanced when the ice cover duration is shortened and the growing season prolonged (Czymzik et al., 2023). In consequence, a prolonged (shortened) ice cover season might also restrict (increase) ¹⁰Be and ⁹Be transport from the catchment into KKJ. In addition, most of the ¹⁰Be record from KKJ sediment reveals amplitude variations of only ~20–30% (Figure 4). These variations do not match the expected ~50% ¹⁰Be production rate changes connected with centennial-scale solar variability (Muscheler and Heikkilä, 2011).

Nevertheless, covariance between the ¹⁴C production time series, as well as ¹⁰Be/⁹Be and TOC corrected ¹⁰Be records from KKJ sediments for most of the investigated time-span point to a removal of the main environmental bias (Figure 6). Some remaining inconsistencies might point to further unresolved environmental influences and advise against the interpretation of small-scale features in terms of cosmogenic radionuclide production rate changes (Muscheler et al., 2014) (Figure 6).

Figure 6.

Synchronization of ¹⁰Be/⁹Be and TOC corrected ¹⁰Be records from Lake Kälksjön (KKJ) sediments with ¹⁴C production rates from the IntCal20 calibration curve (Reimer et al., 2020). (a) ¹⁰Be/⁹Be record (blue) synchronized to ¹⁴C production rates (red). The dashed line depicts the ¹⁰Be/⁹Be record on its original time-scale. The significance of the correlation coefficient takes into account the effects of autocorrelation in the time-series (Ebisuzaki, 1997). (b) Chronological changes applied to the ¹⁰Be/⁹Be record during the synchronization with root mean square error (RMSE) estimates (Grant et al., 2012). (c) Moving correlation between the synchronized ¹⁰Be/⁹Be and ¹⁴C production records. (d) Same as (a), but for the TOC corrected ¹⁰Be record from KKJ sediments. (e) Same as (b), but for the TOC corrected ¹⁰Be record from KKJ sediments. (f) Same as (c), but for the TOC corrected ¹⁰Be record from KKJ sediments.

Synchronizing the WGB and KKJ sediment records to the IntCal20 ¹⁴C time-scale

Shared variability in ¹⁰Be and ¹⁴C records can be interpreted in terms of common changes in cosmogenic radionuclide production and serves as a tool for the synchronization of different environmental archives based on curve fitting (Czymzik et al., 2020; Mekhaldi et al., 2020; Mellström et al., 2015; Snowball et al., 2010). Semi-automatic alignment using the best fit was applied to synchronize the environment-corrected ¹⁰Be records from WGB (¹⁰Be/⁹Be) and KKJ (¹⁰Be/⁹Be and ¹⁰Be_TOC) sediments to the ¹⁴C production time series from the IntCal20 calibration curve (Figures 5 and 6).

The synchronized ¹⁰Be/⁹Be record from WGB sediments reveals a significant positive correlation with the ¹⁴C production time series (r = 0.54, p < 0.01) (Figure 5). Moving correlations with a window-size of 250 years indicate correlation coefficients between 0.57 and 0.81 from 5200 to 6000 a BP and, as expected, reduced correlations between 0.13 and 0.66 for the part of the WGB ¹⁰Be/⁹Be record from non-laminated sediments with a smoothed ¹⁰Be production signal (Figure 5). Compared to the ¹⁴C production time-series from IntCal20, the WGB ¹⁰Be/⁹Be record before the synchronization appears up to 51 years younger between 5300 and 5700 a BP and up to 33 years older from 6000 to 6200 a BP (Figure 5). The applied chronological changes are outside the ±21–24 years RMSE uncertainties for the synchronization procedure for most of the record, but within the centennial chronological uncertainties of the ¹⁴C dated WGB record (Figure 5) (Warden et al., 2017).

The synchronized ¹⁰Be/⁹Be and ¹⁰Be_TOC records from KKJ sediments reveal significant positive correlations with the ¹⁴C production time series (Figure 6). Moving correlations vary between 0.39 and 0.94 for the ¹⁰Be/⁹Be-ratio, as well as 0.4 and 0.87 for the ¹⁰Be_TOC record. Exceptions are short intervals with correlation coefficients down to <0.2 at ~6050 and 5700 a BP for the ¹⁰Be/⁹Be, as well as ~5800 a BP for the ¹⁰Be_TOC record corresponding to short-term mismatches possibly associated with secondary ¹⁴C peaks that are not resolved in the KKJ ¹⁰Be record at ~30-year resolution (Figure 6).

The chronological refinements suggested by the applied synchronizations for KKJ sediments are similar for both the ¹⁰Be/⁹Be-ratio and ¹⁰Be_TOC records (Figure 6). They suggest that the KKJ chronology is about 20 years too old between 5300 and 6000 a BP and 40 years too young in the interval from 6200 to 6400 a BP (Figure 6). All performed chronological changes occur within the RMSE uncertainties of the applied synchronization procedure (¹⁰Be/⁹Be: ±30-41 years; ¹⁰Be_TOC: ±30-42 years) and ¹⁴C-based chronology of the KKJ sediment core (Czymzik et al., 2023). The resulting uncertainty ranges for the here applied synchronization procedure are slightly larger than the ±20 years from ¹⁴C wiggle-match dating of KKJ sediments from 3000 to 2000 a BP, likely due to our continuous sampling strategy that was not adjusted to changes in the ¹⁴C calibration curve (Mellström et al., 2015).

Baltic Sea ventilation ~6250 to 6000 a BP

Oxygen levels and sediment deposition in the deep basins of the central Baltic Sea before the time of major human interferences were predominantly modulated by coinciding centennial climate periods (Andrén et al., 2000; Moros et al., 2020; Zillén et al., 2008). Warmer periods like the Holocene Thermal Maximum (HTM; ~8000 to 5000 a BP), Roman Period (~2200 to 1500 a BP) and Medieval Climate Anomaly (~1300 to 800 a BP) were accompanied by strengthened water column stratification and the preservation of finely laminated sediments under anoxic conditions (Moros et al., 2020; Wanner et al., 2011). Colder periods during the Neoglaciation (~4000 to 2200 a BP) and Little Ice Age (~500 to 100 a BP) are paralleled by water column ventilation and lateral sediment transport causing the deposition of homogenous sediments in the deep Baltic Sea basins (Moros et al., 2020; Zillén et al., 2008). A proposed mechanism is deep water formation during colder winters (Moros et al., 2020). One seemingly exception to this general pattern is a period of homogenous sedimentation in the central Baltic Sea basins between ~6250 and 6000 a BP, intercalated into the finely laminated sediments of the HTM (Warden et al., 2017) (Figure 7). This period is broadly coincident with a shorter cold period predominantly in the southern Hemisphere, North America and Greenland (Lecavalier et al., 2017; Wanner et al., 2011) and intercalated into intervals of distinct stratification in the Bothnian Sea (Kaiser et al., 2024).

Figure 7.

Comparison of synchronized paleoenvironmental records from Baltic Sea and Lake Kälksjön (KKJ) sediments. (a) Total organic carbon (TOC) record from Baltic Sea sediment core P435-2-1 (Warden et al., 2017). (b) TEX₈₆ record from Baltic Sea sediment core P435-2-1 interpreted to reflected subsurface water temperatures at 80–120 m depth (Warden et al., 2017; Wittenborn et al., 2022). (c) TOC record from Lake Kälksjön (KKJ) sediments interpreted to reflect North Atlantic Oscillation (NAO) driven changes in ice cover and winter temperatures (Czymzik et al., 2023). All records were synchronized to the IntCal20 ¹⁴C time scale based on the existing ¹⁰Be data from both archives (~6400-5200 a BP). Within the older and younger parts bracketing this interval, age offsets from the start and end-points of the synchronization were extrapolated. The gray bar highlights the discussed cold reversal within the HTM resulting in homogenous sediments deposition in the central Baltic Sea basins.

Comparing the occurrence of this homogenous sediment sequence with temperature proxies from the synchronized Baltic Sea and KKJ sediments covering most of the HTM, allows us to investigate the possible triggering mechanisms, with high chronological accuracy. Before the comparison, the here applied chronological refinements of WGB sediment core M86-1a/33-4GC were transferred to Eastern Gotland Basin sediment core P435-2-1 providing a TEX₈₆ biomarker record (Warden et al., 2017; Wittenborn et al., 2022). Downcore calibration studies indicate this record to reflect central Baltic Sea water temperatures at 80–120 m depth, controlled by inter-annual to decadal changes in regional air temperature (Meier, 2005; Wittenborn et al., 2022).

The interval of homogenous sedimentation is paralleled by reduced inter-annual air temperatures in the central Baltic Sea region, reflected by the TEX₈₆ record (Figure 7) (Wittenborn et al., 2022). Moreover, this interval coincides with colder winters, imprinted in reduced TOC contents of KKJ sediments (Czymzik et al., 2023). Periods of reduced TOC contents in KKJ sediments were interpreted to reflect colder winters reducing growth season length during a prevailing negative mode of the North Atlantic Oscillation (NAO) (for details see Czymzik et al., 2023). Its association with changes in NAO polarity supports a central Baltic Sea basin wide geographic relevance of the recorded TOC signal in KKJ sediments (Czymzik et al., 2023). However, cooler temperatures during this period were not unique and comparably cold intervals occurred from 6800 to 6500 a BP in the Baltic Sea record, and around 5400 and 6450 a BP in the KKJ data (Figure 7).

The perhaps exceptional feature during the period of homogenous sediment deposition from 6250 to 6000 a BP is the coinciding decrease in both, inter-annual (Baltic Sea) and winter (KKJ) temperatures (Figure 7). Therefore, we speculate that such inter-annual cooling might have been prerequisite to sufficiently reduce vertical water temperature gradients and allow deep water formation in the central Baltic Sea basins during exceptionally cold winters, also within the generally warmer HTM conditions (Moros et al., 2020). The slight temporal lag of the central Baltic Sea TEX₈₆ temperature record compared to the winter temperature time series from KKJ sediments might reflect a delay of ~10 years required for the atmospheric temperature signal to propagate into the deeper Baltic Sea water column, where it is transferred into the biomarker temperature proxy by biosynthesizing archaea (Figure 7) (Meier, 2005; Wittenborn et al., 2022).

Conclusions

We present two new ¹⁰Be records from brackish WGB and lacustrine KKJ sediments covering the time interval from ~6400 to 5200 a BP. ¹⁰Be/⁹Be-ratios from both archives, as well as regression analysis with the correlating TOC record from KKJ sediments, reduced non-production signals in the measured ¹⁰Be records.

Significant correlations between the ¹⁰Be and ¹⁰Be/⁹Be records from WGB sediments with the ¹⁴C production rate time series from the IntCal20 calibration curve indicate that environmental influences on ¹⁰Be concentrations in the investigated sediment section are minor. The main exception is a ~4 cm thick non-laminated sediment unit without decadal-scale ¹⁰Be production variations, likely smoothed out by turbulent sediment mixing. Similarities between the ¹⁰Be, ⁹Be and TOC records from KKJ sediments, in turn, point to a climate influence on ¹⁰Be deposition in this archive. However, significant correlations of the¹⁰Be/⁹Be and ¹⁰Be_TOC records with the ¹⁴C production time series point to a broad removal of this bias.

The preserved cosmogenic radionuclide production variations allowed us to synchronize the ¹⁰Be/⁹Be-ratios from WGB and KKJ sediments, as well as the ¹⁰Be_TOC record from KKJ sediments to the ¹⁴C production time series via curve fitting and reduce the centennial chronological uncertainties to about ±20 (WGB) and ±40 (KKJ) years. Integrating synchronized proxy records from both archives enabled us to interpret a relatively short period of homogenous sediment deposition and ventilation in the central Baltic Sea basins from 6250 to 6000 a BP as possibly triggered by a coinciding mean annual and winter cooling allowing deep water formation during the generally warmer HTM. Such robust multi-archive investigations of short climate events would not have been possible based on the originally centennial-scale uncertainties of the ¹⁴C dated sediment records. Future studies with longer ¹⁰Be time series could further improve the robustness of the combined Baltic Sea and KKJ sediment chronologies and test our paleoclimate implications.

Footnotes

Acknowledgements

We thank two anonymous reviewers and the editor for their valuable comments. We are grateful to Brian Brademann (GFZ) and Hendrick Mück (IOW) for their help during the coring campaign at Lake Kälksjön, as well as Anne Köhler and Alma Schulz (IOW) for lab assistance. ¹⁰Be and ⁹Be data measured in this study will be published in the PANGAEA data library. Holocene ¹⁴C production rates inferred from the IntCal20 calibration curve will be made available upon request by RM.

Author contributions

Markus Czymzik: Designed the study; led the extraction of ¹⁰Be from sediment samples; performed the statistical analyses and wrote the initial draft.

Marcus Christl: Conducted ¹⁰Be measurements. Contributed to the interpretation of the data and the writing of the final manuscript.

Olaf Dellwig: Performed ⁹Be measurements. Contributed to the interpretation of the data and the writing of the final manuscript.

Raimund Muscheler: Contributed to the interpretation of the data and the writing of the final manuscript.

Daniela Müller: Contributed to the interpretation of the data and the writing of the final manuscript.

Jérôme Kaiser: Contributed to the interpretation of the data and the writing of the final manuscript.

Markus J Schwab: Contributed to the interpretation of the data and the writing of the final manuscript.

Carla KM Nantke: Contributed to the interpretation of the data and the writing of the final manuscript.

Achim Brauer: Contributed to the interpretation of the data and the writing of the final manuscript.

Helge W Arz: Led XRF and LOI measurements. Contributed to the interpretation of the data and the writing of the final manuscript.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: M.Cz. is financed through grants CZ 227/4-1 (SyncBalt project) and CZ 227/7-1 (BaltChron project) of the German Science Foundation (DFG). This publication is a contribution to the BaltRap project SAW-2017-IOW-2, funded by the Leibniz Association.

ORCID iDs

Markus Czymzik

Marcus Christl

Jérôme Kaiser

References

Adolphi

Muscheler

(2016) Synchronizing the Greenland ice core and radiocarbon timescales over the Holocene – Bayesian wiggle-matching of cosmogenic radionuclide records. Climate of the Past 12: 15–30.

Aldahan

Possnert

(1998) A high-resolution ¹⁰Be profile from deep sea sediment covering the last 70ka: Indication for globally synchronized environmental events. Quaternary Geochronology 17: 1023–1032.

Andrén

Kunzendorf

(2000) Holocene history of the Baltic Sea as a background for assessing records of human impact in the sediments of the Gotland Basin. Holocene 10: 687–702.

BACC (2015) Second Assessment of Climate Change for the Baltic Sea Basin. Heidelberg: Springer.

Balco

DeJong

Ridge

, et al. (2021) Atmospherically produced beryllium-10 in annually laminated late-glacial sediments of the North American Varve Chronology. Geochronology 3: 1–33.

Berggren

Aldahan

Possnert

, et al. (2010) ¹⁰Be and solar activity cycles in varved lake sediments, AD 1900–2006. Journal of Paleolimnology 44: 559–569.

Berglund

Sandgren

Barnekow

, et al. (2005) Early Holocene history of the Baltic Sea, as reflected in coastal sediments in Blekinge, southeastern Sweden. Quaternary International 130: 111–139.

Björck

(1995) A review of the history of the Baltic Sea, 13.0-8.0 ka BP. Quaternary International 27: 19–40.

Blockley

SPE

Bourne

Brauer

, et al. (2014) Tephrochronology and the extended intimate (integration of ice-core, marine and terrestrial records) event stratigraphy 8–128 ka b2k. Quaternary Science Reviews 106: 88–100.

10.

Christl

Vockenhuber

Kubik

, et al. (2013) The ETH Zurich AMS facilities: Performance parameters and reference materials. Nuclear Instruments & Methods in Physics Research Section B, Beam Interactions With Materials and Atoms 294: 29–38.

11.

Croudace

Löwemark

Tjallingii

, et al. (2019) Current perspectives on the capabilities of high resolution XRF core scanners. Quaternary International 514: 5–15.

12.

Czymzik

Adolphi

Muscheler

, et al. (2016) A varved lake sediment record of the ¹⁰Be solar activity proxy for the Lateglacial-Holocene transition. Quaternary Science Reviews 153: 31–39.

13.

Czymzik

Muscheler

Adolphi

, et al. (2018) Synchronizing ¹⁰Be in two varved lake sediment records to IntCal13 ¹⁴C during three grand solar minima. Climate of the Past 14: 687–696.

14.

Czymzik

Muscheler

Brauer

, et al. (2015) Solar cycles and depositional processes in annual ¹⁰Be from two varved lake sediment records. Earth and Planetary Science Letters 428: 44–51.

15.

Czymzik

Nowaczyk

Dellwig

, et al. (2020) Lagged atmospheric circulation response in the Black Sea region to Greenland interstadial 10. Proceedings of the National Academy of Sciences of the United States of America 117: 28649–28654.

16.

Czymzik

Tjallingii

Plessen

, et al. (2023) Mid-Holocene reinforcement of North Atlantic atmospheric circulation variability from a western Baltic lake sediment record. Climate of the Past 19: 233–248.

17.

Dellwig

Wegwerth

Schnetger

, et al. (2019) Dissimilar behaviors of the geochemical twins W and Mo in hypoxic-euxinic marine basins. Earth-Science Reviews 193: 1–23.

18.

Ebisuzaki

(1997) A method to estimate the statistical significance of a correlation when the data are serially correlated. Journal of Climate 10: 2147–2153.

19.

Ekman

(1996) A consistent map of the postglacial uplift of Fennoscandia. Terra Nova 8: 158–165.

20.

Emeis

Struck

Blanz

, et al. (2003) Salinity changes in the central Baltic Sea (NW Europe) over the last 10000 years. Holocene 13: 411–421.

21.

Fleitmann

Cheng

Badertscher

, et al. (2009) Timing and climatic impact of Greenland interstadials recorded in stalagmites from northern Turkey. Geophysical Research Letters 36: L19707.

22.

Grant

Rohling

Bar-Matthews

, et al. (2012) Rapid coupling between ice volume and polar temperature over the past 150,000 years. Nature 491: 744–747.

23.

Häusler

Moros

Wacker

, et al. (2017) Mid- to late Holocene environmental separation of the northern and central Baltic Sea basins in response to differential land uplift. Boreas 46: 111–128.

24.

Heaton

Blaauw

Blackwell

, et al. (2020) The IntCal20 approach to radiocarbon calibration curve construction: A new methodology using Bayesian splines and errors-in-variables. Radiocarbon 62: 821–863.

25.

Henken-Mellies

Beer

Heller

, et al. (1990) ¹⁰Be and ⁹Be in South Atlantic DSDP Site 519: Relation to geomagnetic reversals and to sediment composition. Earth and Planetary Science Letters 98: 267–276.

26.

Jokinen

Virtasalo

Jilbert

, et al. (2018) A 1500-year multiproxy record of coastal hypoxia from the northern Baltic Sea indicates unprecedented deoxygenation over the 20th century. Biogeosciences 15: 3975–4001.

27.

Kaiser

Tomczak

Dellwig

, et al. (2024) Mediterranean-like “fall dump” events in the Baltic Sea. Holocene 34: 415–419.

28.

Kaiser

van der Meer

MTJ

Arz

(2017) Long-chain alkenones in Baltic Sea surface sediments: New insights. Organic Geochemistry 112: 93–104.

29.

Lal

Peters

(1967) Cosmic ray produced radioactivity on the Earth. In: Flügge

(ed.) Handbuch Der Physik. Berlin: Springer, pp.551–612.

30.

Lane

Brauer

Martín-Puertas

, et al. (2015) The Late Quaternary tephrostratigraphy of annually laminated sediments from Meerfelder Maar, Germany. Quaternary Science Reviews 122: 192–206.

31.

Lecavalier

Fisher

Milne

, et al. (2017) High Arctic Holocene temperature record from the Agassiz ice cap and Greenland ice sheet evolution. Proceedings of the National Academy of Sciences of the United States of America 114: 5952–5957.

32.

Leppäranta

Myrberg

(2009) Physical Oceanography of the Baltic Sea. Berlin, Heidelberg: Springer.

33.

Lisiecki

(2002) Application of dynamic programming to the correlation of paleoclimate records. Paleoceanography 17: 1049.

34.

Meier

HEM

(2005) Modeling the age of Baltic seawater masses: Quantification and steady state sensitivity experiments. Journal of Geophysical Research Oceans 110: C02006.

35.

Mekhaldi

Czymzik

Adolphi

, et al. (2020) Radionuclide wiggle matching reveals a nonsynchronous early Holocene climate oscillation in Greenland and western Europe around a grand solar minimum. Climate of the Past 16: 1145–1157.

36.

Mellström

Nilsson

Stanton

, et al. (2015) Post-depositional remanent magnetization lock-in depth in precisely dated varved sediments assessed by archaeomagnetic field models. Earth and Planetary Science Letters 410: 186–196.

37.

Moros

Kotilainen

Snowball

, et al. (2020) Is ‘deep-water formation’ in the Baltic Sea a key to understanding seabed dynamics and ventilation changes over the past 7,000 years? Quaternary International 550: 55–65.

38.

Muscheler

Adolphi

Knudsen

(2014) Assessing the differences between the IntCal and Greenland ice-core time scales for the last 14,000 years via the common cosmogenic radionuclide variations. Quaternary Science Reviews 106: 81–87.

39.

Muscheler

Beer

Kubik

, et al. (2005) Geomagnetic field intensity during the last 60,000 years based on ¹⁰Be and ³⁶Cl from the summit ice cores and ¹⁴C. Quaternary Science Reviews 24: 1849–1860.

40.

Muscheler

Heikkilä

(2011) Constraints on long-term changes in solar activity from the range of variability of cosmogenic radionuclide records. Astrophysics and Space Sciences Transactions 7: 355–364.

41.

Raisbeck

Yiou

Fruneau

, et al. (1981) Cosmogenic ¹⁰Be/⁷Be as a probe of atmospheric transport processes. Geophysical Research Letters 8: 1015–1018.

42.

Rasmussen

Bigler

Blockley

, et al. (2014) A stratigraphic framework for abrupt climatic changes during the last glacial period based on three synchronized Greenland ice-core records: Refining and extending the INTIMATE event stratigraphy. Quaternary Science Reviews 106: 14–28.

43.

Reimer

Austin

WEN

Bard

, et al. (2020) The IntCal20 Northern Hemisphere radiocarbon age calibration curve (0–55 cal kBP). Radiocarbon 62: 725–757.

44.

Savranskaia

Egli

Valet

J-P

, et al. (2021) Disentangling magnetic and environmental signatures of sedimentary ¹⁰Be/⁹Be records. Quaternary Science Reviews 257: 106809.

45.

Siegenthaler

(1983) Uptake of excess CO₂ by an outcrop-diffusion model of the ocean. Journal of Geophysical Research 88: 3599–3608.

46.

Simon

Thouveny

Bourlès

, et al. (2016) Authigenic ¹⁰Be/⁹Be ratios and ¹⁰Be-fluxes (230Thxs-normalized) in central Baffin Bay sediments during the last glacial cycle: Paleoenvironmental implications. Quaternary Science Reviews 140: 142–162.

47.

Simon

Thouveny

Bourlès

, et al. (2020) Cosmogenic ¹⁰Be production records reveal dynamics of geomagnetic dipole moment (GDM) over the Laschamp excursion (20–60 ka). Earth and Planetary Science Letters 550: 116547.

48.

Snowball

Muscheler

Zillén

, et al. (2010) Radiocarbon wiggle matching of Swedish lake varves reveals asynchronous climate changes around the 8.2 kyr cold event. Boreas 39: 720–733.

49.

Stanton

Snowball

Zillén

, et al. (2010) Validating a Swedish varve chronology using radiocarbon, palaeomagnetic secular variation, lead pollution history and statistical correlation. Quaternary Geochronology 5: 611–624.

50.

von Grafenstein

Erlenkeuser

Brauer

, et al. (1999) A Mid-European decadal isotope-climate record from 15,500 to 5000 years B.P. Science 284: 1654–1657.

51.

Wanner

Solomina

Grosjean

, et al. (2011) Structure and origin of Holocene cold events. Quaternary Science Reviews 30: 3109–3123.

52.

Warden

Moros

Neumann

, et al. (2017) Climate induced human demographic and cultural change in northern Europe during the mid-Holocene. Scientific Reports 7: 15251.

53.

Winterhalter

(1992) Late-Quaternary stratigraphy of Baltic Sea basins - a review. Bulletin - Geological Society of Finland 64: 189–194.

54.

Wittenborn

Radtke

Dutheil

, et al. (2022) A downcore calibration of the TEX₈₆^L te mperature proxy for the Baltic Sea. Continental Shelf Research 251: 104875.

55.

Wittmann

von Blanckenburg

Bouchez

, et al. (2012) The dependence of meteoric ¹⁰Be concentrations on particle size in Amazon river bed sediment and the extraction of reactive ¹⁰Be/⁹Be ratios. Chemical Geology 318-319: 126–138.

56.

Wulf

Ott

Słowiński

, et al. (2013) Tracing the Laacher See tephra in the varved sediment record of the Trzechowskie palaeolake in central northern Poland. Quaternary Science Reviews 76: 129–139.

57.

Ziegler

Jilbert

de Lange

, et al. (2008) Bromine counts from XRF scanning as an estimate of the marine organic carbon content of sediment cores. Geochemistry Geophysics Geosystems 9: Q05009.

58.

Zillén

Conley

Andrén

, et al. (2008) Past occurrences of hypoxia in the Baltic Sea and the role of climate variability, environmental change and human impact. Earth-Science Reviews 91: 77–92.

Synchronizing the Western Gotland Basin (Baltic Sea) and Lake Kälksjön (central Sweden) sediment records using common cosmogenic radionuclide production variations

Abstract

Keywords

Introduction

Study sites

Material and methods

Sediment cores and initial chronologies

Beryllium extraction and measurements

XRF element profiles and loss on ignition

Correcting for environmental influences on 10Be concentrations

14C production rates

Time scale synchronization and propagating uncertainties

Results

Beryllium inventories and sediment composition

Discussion

Mechanisms of 10Be deposition in WGB and KKJ sediments

Synchronizing the WGB and KKJ sediment records to the IntCal20 14C time-scale

Baltic Sea ventilation ~6250 to 6000 a BP

Conclusions

Footnotes

Acknowledgements

Author contributions

Funding

ORCID iDs

References

Correcting for environmental influences on ¹⁰Be concentrations

¹⁴C production rates

Mechanisms of ¹⁰Be deposition in WGB and KKJ sediments

Synchronizing the WGB and KKJ sediment records to the IntCal20 ¹⁴C time-scale