Characterization of (paleo)lacustrine landforms using sedimentological and portable OSL investigations at Schweriner See, north‐eastern Germany

We investigated four subaerial (paleo)lacustrine landforms at the north‐eastern shoreline of Schweriner See, north‐eastern Germany. These included two beach ridges, one subaerial nearshore bar and a silting up sequence located close to a fossil cliff, which marks the former maximum extent of Schweriner See. We used luminescence profiling with a SUERC portable OSL device (POSL) on all four sediment sequences in combination with sedimentological investigations such as grain size, loss‐on‐ignition and magnetic susceptibility to provide information on the various formations in a lacustrine depositional environment. The POSL reader was used on pre‐treated polymineral samples to gain an insight into luminescence distribution within the individual sediment sequences, but also among the four sequences. POSL proved valuable to understand depositional processes, which were not visible in lithology or sedimentological parameters. With somewhat larger uncertainty this method provides relative chronologies of the sediment sequences. Additionally, we carried out radiocarbon dating and full optical stimulated luminescence (OSL) dating to establish a chronological framework. OSL ages proved to be more reliable to date beach ridges in this setting than radiocarbon samples, which were severely influenced by sediment reworking. This combined approach of sedimentological analyses, luminescence profiling and absolute age determinations revealed details in depositional processes at Schweriner See which otherwise would have remained undetected. Furthermore, it helped to set these subaerial (paleo)lacustrine landforms in a chronological framework.


| INTRODUCTION
Paleoshorelines include various depositional landforms, which can be used as a paleoenvironmental record. Especially beach ridges are a potential geoarchive for environmental reconstructions worldwide.
They form as depositional landforms at shorelines with shallow shoreface and a minimum amount of accommodation space. Beach ridge development depends on sufficient sediment supply and a strong enough wave energy to deposit predominantly sandy sediment, which results in shoreline progradation (Scheffers et al., 2012;Tamura, 2012). However, beach ridge growth is not only influenced by wave energy and sediment supply, but also by water-level changes, which cause sediment dislocation.
In the past, research on beach ridges has been carried out more frequently in coastal marine settings. However, under favourable conditions beach ridges are also deposited at lake shorelines. Detailed investigations on beach ridges in lake systems have mostly been carried out in large systems such as the Great Lakes, USA (Baedke et al., 2004;Carter, 1986;Fraser & Hester, 1977;Johnston et al., 2007;Petty et al., 1996;Thompson, 1992;Thompson & Baedke, 1995, 1997. However, beach ridge formation is not limited to large lakes. In north-eastern Germany, beach ridges were deposited at shorelines of much smaller lake systems such as Lake Müritz (117 km 2 ; Küster, 2013;Lampe et al., 2009) or Krakower See (17 km 2 ; Lorenz, 2007), but also at small Fürstenseer See (2 km 2 ; Kaiser et al., 2014). Here, we present a set of ridges with different formation mechanisms from the north-eastern shoreline of Schweriner See ( Figure 1, 61.54 km 2 ), which is also located in north-eastern Germany.
To use ridge structures as a paleoenvironmental indicator, a reliable chronology is indispensable (Tamura, 2012). Since beach ridges are mostly composed of siliciclastic material, optically stimulated luminescence (OSL) has proven to be most useful because it can be applied directly to sediment grains to determine burial time (e.g. Dougherty et al., 2019;Preusser et al., 2008;Rhodes, 2011;Tamura, 2012). This method has been widely applied to establish beach ridge chronologies (e.g. Botha et al., 2018;Nott et al., 2009;Preusser et al., 2008;Tamura et al., 2019). In addition to full singlealiquot regenerative OSL (SAR-OSL) dating, the introduction of the SUERC portable OSL reader (Sanderson & Murphy, 2010) proved valuable as a rapid and cost-effective luminescence method to establish relative chronologies and to show inconsistencies in chronologies.
The portable OSL (POSL) method enables the interpretation of chronologies and targeted full SAR-OSL sampling (Stone et al., 2019). Previous studies investigated age structures of sediment sequences (Gray et al., 2018;Stone et al., 2015), soil mixing (Stang et al., 2012) or sedimentological processes (Muñoz-Salinas et al., 2012). However, the luminescence signal obtained with the POSL reader is not only a function of post-depositional age, but also of luminescence sensitivity, dose rates and signal resetting (Sanderson & Murphy, 2010). Therefore, parameters such as moisture, organic matter content, mineralogy and the geological background have to be considered for the interpretation ). An extensive overview and state of the science is given by Munyikwa et al. (2021).
In this study we explore the potential of a combined approach of luminescence profiling with sedimentological analyses to characterize internal sedimentary structures of (paleo)lacustrine landforms found at the north-eastern shoreline of Schweriner See. We investigated F I G U R E 1 (a) Location of Schweriner See in north-eastern Germany (inset) and overview of Schweriner See and surrounding lakes using a digital terrain model in 5 m resolution including water depth. The area of investigation is shown in a bold rectangle. (b) The investigation area 'Buerwischen' is located at the north-eastern part of Schweriner See (square in a). Geomorphological features at the sampling location were investigated using sediment cores as well as soil pits (coloured dots). Dashed lines highlight the width of the ridges. (c) Transect x-x 0 shows the elevation of the investigated subaerial ridges extracted from the DEM as well as locations of cores and pits [Color figure can be viewed at wileyonlinelibrary.com] two beach ridges, one subaerial nearshore bar and one silting-up sediment sequence in front of a fossil cliff which marks the former maximum lake extent. Geochronological data from radiocarbon and full SAR-OSL dating will furthermore support the interpretation and form a chronological framework.

| SITE DESCRIPTION
Schweriner See (53 43.256 0 N, 11 27.544 0 E, Figure 1) is located at $37.8 m a.s.l. in the western part of the Mecklenburg lake district (Mecklenburgische Seenplatte), north-eastern Germany, approximately 20 km south of the Baltic Sea. The lake has a surface area of 61.54 km 2 and extends over 24.8 km in the N-S and up to 6 km in the E-W direction. The term 'Schweriner See' refers to two subbasins, which are separated by an (in parts) artificial dam that was completed through a swampy surface (Ramper Moor, Figure 1) to connect the western and eastern shoreline in AD 1848. The northern part is called Schweriner Außensee while the southern basin is named Schweriner Innensee. Today, water exchange between both basins is ensured by a natural channel. The lake has a complex morphometry with several deep areas and channel structures. Schweriner Außensee also has an extended shallow-water area in front of the eastern shoreline, with water depth < 5 m ( Figure 1). The shoreline of Schweriner Außensee is mostly surrounded by cliffs of fluvioglacial sediment and only a few accumulation areas, where fossil shore landforms such as beach ridges have been developed and preserved.
Schweriner See is situated between two ice-marginal positions (IMPs) of the Weichselian glaciation: the Frankfurt/Brandenburg IMP in the South and the Pomeranian IMP in the North (Figure 1). The Frankfurt/Brandenburg IMP was dated to 19-18 kyr (Heine et al., 2009) and for the Pomeranian IMP an average age of 16 kyr (Heine et al., 2009;Rinterknecht et al., 2014) was calculated. These ages were dated using cosmogenic 10 Be dating and were later recalibrated to 24-21 and 20-25 ka with an updated global 10 Be production dataset (Hardt & Böse, 2016). The Frankfurt advance formed the lake basin as a tunnel valley, which was later used as a meltwater channel by the melting Pomeranian advance (Krienke & Obst, 2011).
These meltwaters cut through the Frankfurt IMP and created the Stör Valley at the southern end of Schweriner See, which today is the natural outlet of the lake (Figure 1). Outwash sediments from different ice advances are characteristic for the area south of Schweriner See, as well as for the area between the Pomeranian IMP and the lake basin. In between, mainly glaciofluvial sediments, which form plateaus, and ground moraine sediments were deposited and few peatlands emerged (e.g. between both lake basins or at the north-eastern end of Schweriner See) (Krienke & Obst, 2011). The northern outlet, the socalled Wallensteingraben, was built during the 16th century AD through the Pomeranian IMP to connect Schweriner See with the Baltic Sea. The additional drainage most likely led to a lake-level lowering, which has not been quantified until today (von Carmer, 2006). Nowadays, Schweriner See is mainly fed by groundwater, precipitation and surrounding waters, which act as tributaries (Nixdorf et al., 2004).
The recent climate of the study area is documented by the Deutscher Wetterdienst (DWD, German Meteorological Institute).
The research area has a mean annual temperature of 9.5 C with the coldest month being January (1.2 C) and the warmest July (18.4 C) (1991DWD Climate Data Centre, 2021a). Mean annual precipitation is 645 mm with dominating summer rainfalls (1991DWD Climate Data Centre, 2021b). However, 2018 was exceptionally dry (442 mm) and warm (10.6 C), which led to a significant drop in lake level, exceeding the minimum operating level. The main wind direction is SSW (194 C, 1964SSW (194 C, -2019  The term BUW plus a number (10,12,17,19/2) will be used in the following to distinguish between the different profiles. Profiles

| MATERIALS AND METHODS
Subaerial ridge structures were identified using a DEM5 with a 5 Â 5 m resolution before fieldwork ( Figure 1b). Positions were recorded with a handheld GPS. Altitudes were determined using the DEM5. Numbers refer to profiles and sediment cores at the same time. Identified key positions of (paleo)lacustrine landforms (BUW10, BUW12, BUW17, and BUW19/2) were studied in-depth. All positions were investigated using soil pits and sediment cores. Photographs, detailed lithological descriptions (grain size, colour, moisture content, percentage of organic matter, carbonate content) following Ad-hoc-AG Boden (2005) and sediment samples for OSL were obtained from the pits. Samples for full SAR-OSL dating and dose rate determination were sampled from each lithological unit in opaque tubes. Additional samples for radiocarbon dating were taken from organic-rich layers (e.g. peat or paleosols) to provide an independent age control. Sediment cores were taken in opaque plastic liners 1 m behind the pit.
They were investigated with sedimentological methods and luminescence profiling in the Physical Geography Laboratory of the University of Greifswald. The approach is shown in Figure S1.
Before subsampling, magnetic susceptibility was measured on all sediment cores (Bartington MS2C sensor, diameter 80 mm). Subsequently, all sediment cores were opened under red-light conditions. We pursued luminescence profiling and sedimentological methods (water content and dry bulk density determination, loss on ignition [LOI] as estimate for the organic matter content [OM] and grain size analysis). For luminescence profiling a pre-treatment protocol (carbonate and organic matter removal, dispersing) similar to the sample preparation for full OSL was applied to reduce the scatter between the samples and to concentrate the targeted quartz and feldspar grains (Gray et al., 2018). The methods are shown in detail in the online Supporting Information Table S1 and   Table S2.
For the luminescence profiling we used a POSL reader by the Scottish Universities Environmental Research Centre (SUERC; Munyikwa et al., 2021;Sanderson & Murphy, 2010). All measurements were performed on dried pre-treated polymineral samples, from which approximately 1-2 g was filled in a 3 cm-diameter aluminium cup. The sample surface in the cups was compressed with a light weight to smooth the surface, ensuring a monolayer as well as a consistent distance to the photomultiplier. Each sample was measured twice using fresh unexposed grains each time. The results of both measurements were averaged. In the case of sediment core BUW12, we merged some neighbouring samples (even though it enlarged the sampling distance), because sample quantities were too small to perform reliable measurements. Infrared (IR) stimulation targets luminescence signals dominated by feldspar. However, Bateman et al. (2015) showed that full depletion is not attained after 60 s IR stimulation due to the limited power of the POSL device and a larger number of grains used in contrast to standard OSL aliquots. Accordingly, the blue-light stimulation (post-IRSL OSL) signal includes a (small) share of IRSL derived by the feldspar signal and cannot be interpreted as pure quartz signal. Nevertheless, an IRSL/OSL ratio was calculated as an approximation to a feldspar/quartz ratio. IRSL and post-IRSL OSL signals were background corrected and depletion indices were calculated. Besides the rapidity of bleaching, the depletion indices of POSL measurements can shed light on grain coatings, with a cleaner grain surface having higher depletion indices if the rest of the parameters are identical (Stone et al., 2019). For the pre-treatment protocol used, the measurement protocol and all calculations, see Table S2 in the online Supporting Information.
Additionally, nine samples were selected for full OSL dating following the standard single-aliquot regenerative (SAR) protocol (Murray & Wintle, 2000, 2003 using a Risø TL/OSL-DA-20 on quartz grains (Table 1). Detailed sample preparation and OSL measurement protocols are shown in Table S3 in the online Supporting Information.
Radionuclide contents (K, Th, U) for dose rate calculations were measured with low-level gamma spectrometry (subsamples of $100 g) at VKTA Rossendorf e.V. (Dresden). The environmental dose rate is crucially affected by the water content in the sediment, presenting a major uncertainty (Preusser et al., 2008). The measured water content is influenced by many factors such as weather, season, vegetation, lake-level changes and artificial drainage canals, which have existed in the investigation area at least since the 1780s. In addition to the sitespecific aspects, a change of climate to wetter or drier periods over time must be considered. To reduce the susceptibility to such factors and the influence of measurement errors, the water content for the samples was modelled based on depth, estimated lake-level fluctuations over time, the assumed drainage through the artificial canals and the resulting groundwater level at the sites (Lampe & Lampe, 2018).
From this model a time-weighted average for the water content was determined. Additionally, we performed a linear regression as described by Stone et al. (2015) using SigmaPlot 14 to check if ages can be derived from the POSL signals. For this we used the POSL signals from the same depth as the OSL samples. We used the D e value instead of the already calculated OSL ages and plotted it against the post-IRSL OSL counts ( Figure 5a).
AMS-14 C dating was carried out as independent age control on 10 samples (Table 2) at the Poznan Radiocarbon Laboratory, Poland.
The dating material consisted of wood (n = 5), bulk soil samples (n = 2), charcoal (n = 2), a water plant seed (n = 1) and a modern water plant (n = 1). Calibration was carried out using Calib 8.20 (Stuiver et al., 2020) with the IntCal20 calibration dataset . To differentiate between radiocarbon and OSL dating in the text and figures, they are indicated by a different notation (e.g. calibrated 14 C age = 720 +145 / À40 cal BP; OSL age = 450 AE 45 a). As radiocarbon ages correspond to the year AD 1950 and OSL ages to the sampling date, which in this case is AD 2019, there is an offset of 69 years between radiocarbon and OSL ages. This offset cannot be neglected in this study because of the very young ages. To ensure comparability between both dating methods, OSL ages were adjusted to radiocarbon ages by subtracting 70 years, equalling cal BP and a. From here on only adjusted ages will be used. Original and recalculated ages are both shown in Table 1.

| Subaerial ridge stratigraphy and relief
The four investigated sedimentary sequences can be subdivided into three depositional types according to their morphology and stratigraphies. The two ridges closest to the lake are represented by profiles  Note: BG refers to samples taken from a pit. If BG in the sample name is missing, sample material is from the sediment core. The dated recent water plant indicates a substantial reservoir effect. All samples were calibrated using Calib 8.20 (Stuiver et al., 2020) and the IntCal20 calibration dataset . Samples printed in bold are considered as reliable.
F I G U R E 2 Photographs of the investigated pits. Highlighted are radiocarbon (cal BP) and OSL ages (a). OSL ages were recalculated to fit radiocarbon ages by subtracting 70 years to ensure a comparability between both dating methods. Ages printed in bold are considered reliable. Dotted lines represent lithological boundaries used in the simplified lithological sketches, which are shown on the right against altitude (m a.s.l.). The two profiles most proximal to the lake, BUW10 and BUW12, show a multi-layered development. In contrast, profile BUW17 is mostly homogenous with thin intercalated gravel layers. Profile BUW19/2 shows a silting-up sequence close to a fossil cliff representing the maximum extent of Schweriner See. See Figure 1

| Luminescence profiling and sedimentological parameters
Sedimentological parameters (water content, LOI, DBD, mean grain size) and POSL data (IRSL and post-IRSL OSL counts, depletion, IRSL/OSL ratio) allow a subdivision of the four sediment sequences into different units, which in general are consistent with lithological boundaries (Figures 3 and 4). The letter after the sequence name indicates the sedimentological unit and the added roman numbering the subdividing portable OSL unit. Organic-rich lithologies generally form one sedimentological unit, whereas clastic components form a second and in profile BUW19/2 a third one. Clastic units of the sedimentological parameters can further be subdivided into POSL units (Figures 3 and 4).
The three ridge structures BUW10, BUW12, and BUW17 consist of calcareous gyttja as well as peat with an overall small grain size, high organic matter content and a low density forming the lowermost unit A at the bottom. For unit A, no POSL measurements were conducted because the sediment yielded not enough clastic material.
Unfortunately, at profile BUW12 the calcareous gyttja at the bottom ( Figure 2) was not recovered in the sediment core used for the sedimentological analyses (Figure 3). In contrast, unit B consists of sandy sediments in all three ridge structures (BUW10, BUW12, and BUW17). Unit B is characterized by a very low organic matter content, a high density and often coarse sand. Peaks in magnetic susceptibility generally support a beginning soil formation in paleosols (Dearing, 1999), identified in the lithology. In profile BUW12 the highest magnetic susceptibility peak correlates with gravelly coarse sand, which showed hints of iron-coated grains. Unit BUW12-B is much more variable than unit B in BUW10 and BUW17. Here grain sizes and sorting vary. Smaller grain sizes correspond to a poorer

| Geochronology
The chronological framework (Figure 2) is based on nine OSL ages (Table 1) from the sandy sediment sequences and 10 AMS-14 C ages (  and 28 AE 7 % (UG-151), but in seven out of nine cases are below 25 %. Relative age errors are between 7.42 and 10.92 %, and therefore within a normal range of OSL age uncertainty (Murray & Olley, 2002).
Sample UG-151 marks an overall exception, with a much lower environmental dose rate in comparison to the other samples (Table 1).
The much lower potassium value ($0.6 % in contrast to $0.95-1.48 %, Table 1) points towards a different sediment source. Furthermore, this sample has a high overdispersion and was sampled in proximity to the underlying peat. Peat naturally has a lower dose rate than the surrounding siliciclastic sediment, which has been formed from igneous or metamorphic rocks. We therefore assume that this age is most likely underestimated.

| OSL and radiocarbon ages
For radiocarbon ages in beach ridges, several possible constraints have to be considered. Previously, radiocarbon samples from beach ridges were taken from the swales in between the ridges, which are often filled with peat. However, ages obtained from swales might partially have a large offset to the ridge formation, as ridge and peat did not necessarily form at the same time (Tamura, 2012). For this study, we sampled paleosols and peat layers within the ridge structures for radiocarbon dating. However, sometimes dating material appears to be severely influenced by material dislocation, resulting in unreliable ages. Therefore, we will mostly rely on OSL ages and relative age structures revealed by luminescence profiling. Even though water-lain sediments can be complicated to date (Preusser et al., 2008), they proved to be more reliable than radiocarbon ages.
The most proximal profile to today's lake shoreline, BUW10, was on wood. Although stratigraphically this age is consistent, we suspect this peat layer to be relocated as a peat layer in the same stratigraphic position has a very similar age of 750 +145 / À65 cal BP in BUW12, which is too old for its stratigraphic position there (Figure 2). In BUW12 this age is above a much younger OSL age, which questions a synsedimentary in-situ development of the peat layer and hence the validity of the two radiocarbon ages. It is very likely that the dating material was reworked and depleted radiocarbon was incorporated into the sediment. The topmost paleosol layer in BUW10 at 38.07 m a.s.l. was dated to 550 +85 / À30 cal BP (Poz-123669) ( Figure 2). Considering the very young OSL age below, this indicates incorporation of radiocarbon-depleted organic material, also leading to a too old age estimate for the paleosol layer.
Luminescence dating in profile BUW12 revealed ages of 320 AE 40 a (UG-154) at 37.85 m a.s.l. and 340 AE 40 a (UG-153) at 38.25 m a.s.l. (Figure 2). Although the lower OSL age is younger, both ages are consistent within the error of the method, which-like in BUW10-indicates a very rapid deposition. The three AMS-14 C ages are all significantly older and arranged in a reverse chronological order with depth ( Figure 2). The only age in stratigraphic order is the youngest age of 530 +35 / À25 cal BP (Poz-122715) at 37.23 m a.s.l., which was dated on a large piece of wood at the transition from gyttja to a reworked peat layer with molluscs and sand in between. The radiocarbon age of 750 +145 / À65 cal BP (Poz-122578) at 37.98 m a.s.l. was obtained from charcoal in a thin peat horizon. As discussed for profile BUW10, the age and composition of this peat layer are similar to a peat layer at 37.45 m a.s.l. in profile BUW10, making a contemporaneous reworking of older material likely. The uppermost radiocarbon age of 1620 +80 / À75 cal BP at 38.48 m a.s.l. (Poz-122724) is based on wood taken from a paleosol, which according to its stratigraphic location is also reworked and depleted in radiocarbon.
The sandy sequence in profile BUW17 was OSL dated to 3010 AE 250 a (UG-158) at 37.50 m a.s.l. and to 3050 AE 230 a (UG-156) at 38.10 m a.s.l. (Figure 2). Considering the error of the dating method as in the two previous ridges BUW10 and BUW12, this indicates a rather rapid deposition. The underlying peat layer formed most likely on-site. The oldest radiocarbon age dates the transition from peat to sand to 4600 +210 / À160 cal BP (Poz-123670) at 37.30 m a.s.l., which was measured on water plant seeds. As a modern water plant was dated to 555 AE 30 BP, indicating a substantial reservoir effect, this might have affected the age of the water plant seed as well. However, a radiocarbon age of 4410 +110 / À120 cal BP (Poz-122579) just above the peat layer at 37.40 m a.s.l. carried out on wood (terrestrial origin) supports the assumption that a reservoir effect is negligible in this case. The age offset between this radiocarbon age at 37.40 m a.s.l. and the OSL age at 37.50 m a.s.l. (3010 AE 250 a) of >1500 years within only 10 cm points to a hiatus. The topmost radiocarbon age is dated on bulk sediment and is located at the transition from organic matter-free sand to recent soil formation processes. The age indicates a subsequent landscape stabilization at 1120 +120 / À110 cal BP (Poz-123671) at 38.52 m a.s.l. The boundary shows indications of ploughing, which could have incorporated younger material. This appears to be very likely, since in all three ridges discussed so far, the deposition in the sandy main part of the ridges occurred very rapidly (also in the lower part of BUW17). If the age of 1120 +120 / À110 cal BP reflects the time of deposition, the sedimentation rate would have decreased dramatically or sediments would have been eroded. However, due to our luminescences profiles discussed earlier, this is rather unlikely as a rapid continuous sedimentation is suggested.
l. (Figure 2). Samples were taken from two separate lacustrine sand layers, which are intercalated by a coarse gravel layer. It is most likely that a lower lake level occurred between the two OSL ages dated in profile BUW19/2. Very likely this low stand is expressed by a hiatus accompanied by the coarse gravel layer. The decomposed peat overlaying the lacustrine sand is dated to 1730 +90 / À100 cal BP implies that in our case the POSL signal seems to correlate mostly with age. But it has to be kept in mind that the regression is only based on eight data points, of which five are of a very young age.
Observed scatter in the regression could result from either anomalous fading, differences in luminescence sensitivity, but also from different modes of deposition such as nearshore bar deposition vs. beach ridge deposition, resulting in incomplete bleaching.

| DISCUSSION
Luminescence profiling is a useful addition to sedimentological analyses and absolute age determinations like OSL or radiocarbon dating.
For the investigation of paleosols and their cover sediments in formerly glaciated areas, Kaiser et al. (2020) remarked that a focus on a larger number of ages for problematic profiles might help to tackle possible constraints. In this study paleosols were unsuitable for radiocarbon dating due to sediment reworking. However, luminescence profiling as a cost-and time-effective method was a tremendous help to disentangle conflicting absolute ages. Luminescence profiling not only offered an indication of relative ages, but also showed changes in sedimentation processes that were not always visible in the sedimen- The luminescence signal in our study seems to be predominantly influenced by an age signal built up after deposition ( Figure 5).
Similarly, a predominantly age-related POSL signal was previously shown in several studies in various settings (e.g. Bateman et al., 2015;Gray et al., 2018;Stone et al., 2015Stone et al., , 2019. This is also true for the sediment sequences at Schweriner See. The higher the POSL signals are, the older the sediments. The relationship between D e values and post-IRSL OSL counts is well represented by a linear regression. However, the POSL signal seems to be partially dependent on grain size in the case of profile BUW12, even though we sieved the sediment to <315 μm to reduce the scatter. In profile BUW12, the highest POSL signals correlate with a peak in grain size and magnetic susceptibility. Bishop et al. (2011) also showed a strong correlation between photon counts and grain size, with increasing grain sizes being less bleached and having a higher inherited luminescence in water-lain sediment. The grain size distribution for all profiles from Schweriner See showed predominantly medium and coarse sand with varying percentages. The fine-sand fraction was mostly <5 %, with a few exceptions. However, in BUW12 the grain size composition shifted to almost 96 % coarse grain sand in some samples. Therefore, the composition within the POSL sample also shifted to overall coarser sediment, which might have been less bleached, resulting in higher luminescence values.
In contrast, it is characteristic for profiles BUW10, BUW12 and BUW17 that the POSL signals of the lowest POSL unit (unit B-I in all three profiles) is lower than the POSL signal in the overlaying sediment. This most likely results from the sample location below the recent lake level and from higher lake levels in the past, which is why a different water content for each sample has to be presumed over time. The water content was corrected in the OSL ages, with an adjusted water content for age calculations (Table 1). However, such a correction was not applied to POSL signals. A varying water content over time might also be a reason for the large scatter obtained in the regression analysis ( Figure 5). Furthermore, the lowest units are in proximity to the underlying peat, which also attenuates radiation, resulting in lower luminescence signals. We did not measure luminescence on the peat sequences, even though in previous studies POSL was successfully measured on organic-rich untreated sediment samples Portenga & Bishop, 2016).
Here the pre-treatment protocol, according to which carbonates and organic matter were removed, was the limiting factor because the peat yielded not enough clastic material to carry out reliable measurements. Like for full SAR-OSL measurements, this method is bound to clastic material, but a greater sample quantity is required .
Particularly prone to changes in luminescence are the two profiles most proximal to today's shoreline. Based on the regression results, age dominates the POSL signal. Therefore, the lower part of profile BUW10 has to be older than profile BUW12, even though profile BUW10 is located closer to the recent shoreline (Figures 1 and 5) and-logically-should be younger. This age structure is also confirmed by the OSL ages, but raises the question how the deposition of these two ridges occurred, as beach ridge structures usually become younger with proximity to the recent shoreline. Based on the POSL signals, the OSL ages and the sedimentology, the following scenario is assumed: before the two foremost beach ridges were deposited, the lake level was significantly lower, which was shown in archaeological investigations (Konze, 2017). A widespread peat formed, as indicated by the peat layer in BUW10 und BUW12. With a rising lake level, the peat was flooded and lacustrine sand (marked with a triangle in Figure 5) was deposited on top of the peat. Due to an earlier flooding of the lower part of BUW10, which is located closer to the recent shoreline, these ages are older (590 AE 70 and 560 AE 50 a), but do not represent beach ridge formation yet-rather the deposition of lacustrine sand. Therefore, these ages indicate a rising lake level. In profile BUW10 the luminescence profile has an inverse structure for units 10-B-I to 10-B-III, which is most likely the result of a lower dose rate (UG-151 with 0.77 AE 0.06 mGy a À1 in contrast to 1.10 AE 0.08 mGy a À1 for the overlaying UG-149,  In contrast to these two beach ridge profiles, profile BUW17 suggests a different mode of deposition. We assume that the ridge was deposited as a nearshore bar, which became subaerial by a lowering lake level (Scheffers et al., 2012;Tanner, 1995). The deposition in BUW17 happened rather rapidly as POSL signals, which have almost no variations, suggest.
Modes of beach ridge deposition can differ between sets of beach ridges in the same system (Goslin & Clemmensen, 2017;Nott et al., 2015;Scheffers et al., 2012). At Schweriner See the different stratigraphies point to two modes of deposition. The foremost two ridge structures with profiles BUW10 and BUW12 were deposited during several events, indicated by the intercalated paleosols and organogenic layers. In contrast, profile BUW17 shows no indication of multi-event development, but rather a continuous sedimentation as suggested in the POSL and sedimentological values, which have almost no variations. This hints at either more constant depositional conditions or even a deposition somewhat offshore not directly at the shoreline, which might be more susceptible to changes in grain size due to storm events with varying intensities, wind speeds and directions. Inclining gravel layers suggest a consistent wave direction. A conceivable scenario might be a subaquatic deposition as a nearshore bar, that was exposed by a (rapid?) lake-level decline. The formation of such bars can also be observed today on the shallow water subaquatic extension of the plain. In contrast to the profiles from the ridge structures, profile BUW19/2 represents lacustrine deposits at the former maximum extent of Schweriner See. The lacustrine deposits are covered by peat, which formed after the final lake-level decline and was subsequently buried by colluvial deposits from the adjacent cliff.

| CONCLUSION
Luminescence profiling proved very valuable to further subdivide sed-

DATA AVAILABILITY STATEMENT
The datasets generated and analysed during the current study are available from the corresponding author on reasonable request.