The foraminiferal record in the Holocene evolution of the Mecklenburg Bay (south-western Baltic Sea)

The sediment for this study was obtained from Core 317980-3 collected in the central part of MB (Fig. 1) with a gravity corer during the RV Maria S. Merian cruise in 2006. The coring site (54°12.01′N; 11°21.010′E) was located at a depth of 21.8 m.

Upon retrieval, the core was divided into 1-m long sections and stored under refrigeration. A 0–620 cm core section was used for the analysis of foraminifera. Prior to subsampling and analyses, the core was visually inspected and four lithological units were distinguished. The lowest unit (620–615 cm) consisted of light grey sandy clay. It was overlain by a 10-cm thick layer of brownish-black peat gyttja separated by a sharp boundary from the bottom layer. The peat gyttja was covered by a layer (605–575 cm) of brown-grey sandy silt with a more gradational contact. The top part of the core (575–0 cm) consisted of sand-containing mud, its colour changing visibly from dark grey to light olive-grey and the sand content decreasing upwards. Small shells of the bivalves Limecola (formerly Macoma) sp. and Mytilus sp. were found at depths of 509–495, 462–450, 390 and 50 cm. The uppermost 10-cm core layer was dark grey in colour and exuded a typical H₂S smell (Kostecki et al. 2015). Other properties of the core are summarized in Kostecki et al. (2015).

The core was cut into 1-cm sections from which subsamples were retrieved for lithological, geochemical and diatomological analyses (summarized by Kostecki et al. 2015). As described in Kostecki et al. (2015), the bivalve shells found (Limecola sp. and Mytilus sp.) as well as the bulk sediment (peat gyttja) were used for radiocarbon dating conducted at the Poznań Radiocarbon Laboratory, using the ¹⁴C accelerator mass spectrometry (AMS). Dates from the lacustrine deposits were calibrated using appropriate software (CALIB 8.1 radiocarbon calibration programme – Stuiver & Reimer 1993 as well as Marine20 and IntCal20 – Heaton et al. 2020 and Reimer et al. 2020). The calibrated dates were used to develop an age-depth model for Core 317980-3.

Sediment portions for the foraminiferal analysis were collected at 5-cm intervals. These samples were placed in plastic bags and stored in the refrigerator (4°C) until the analysis.

The sediment portion to be analyzed was placed in a 1-dm³ beaker, covered with 0.6 dm³ water and left for 24 h in the refrigerator (7°C). Subsequently, the sediment was stirred with a glass rod to disintegrate clumps and the beaker content was passed through a set of sieves (mesh sizes: 0.5; 0.250; 0.180; 0.090 and 0.063 mm). The residue from each sieve was rinsed onto Petri dishes, covered with water and examined under a stereomicroscope. Any foraminifera found were identified and counted. As the calcareous (rotaliids) and some arenaceous (trochamminids) foraminifera found in finer sediment fractions (sieving residues on 0.063 and 0.090 mm mesh size sieves) were seldom intact (see below), and were represented mostly by fragments, a foraminifera individual was considered as such and counted whenever the fragment found consisted of at least one whorl. Whorl fragments were disregarded. Whenever the state of preservation allowed, the foraminifera were identified to at least the genus level (consulting inter alia Loeblich & Tappan 1988; Murray & Alve 2011). The foraminiferal nomenclature was checked against the World Register of Marine Species database (WoRMS; www.marinespecies.org). Upon completion of the analysis of a sample, the sieving residues were combined, dried (105°C) for 24 h and weighed on a precision laboratory balance (Ohaus). The foraminiferal abundance is referred to as the number of individuals per gram dry sediment (g d.s.). As the finest (< 0.063 mm) sediment fraction was lost on sieving, the “sediment” is understood as the sum of sediment fractions coarser than 0.063 mm.

Examination of the sieving residues allowed to complement the visual inspection of the whole core (see above) with some details observed under the microscope (e.g. the presence of plant and invertebrate remains such as ostracod valves, variable proportion of sand etc.; cf. Table 1).

Associations between the sedimentary environmental variables described in an earlier paper (Kostecki et al. 2015) and abundances of the dominant foraminifera groups determined in this study were analyzed using the Canonical Correlation Analysis (CCA). In our case, CCA was used to look for linear combinations of environmental variables (the X variables: mean grain size as well as the contents of terrigenous and biogenic silica, magnesium and calcium determined in 126 samples) which produced the best correlations with linear combinations of the groups of foraminiferal taxa (the Y variables: the Ammonia group, Eggerelloides scaber, Trochamminacea, Reophax sp.). CCA was performed using R CRAN packages: CCA (González et al. 2008; González & Déjean 2009) and CCP (Menzel 2009) freely available in the R Archive Network. The data were transformed using the log (x + 1) function to normalize their distributions. The relationships between the environment parameters and foraminifera were inferred from canonical dimensions visualized on the ordination diagram of the most significant CCA coefficients.

The calcareous (rotaliid) foraminifera present in the sediment were mostly decalcified and occurred as organic shell linings (Fig. 2) or fragments thereof. The decalcified forms were identified as representing mainly two species of Ammonia (see below). The agglutinated forms, on the other hand, were usually well preserved, although the finer sediment fractions contained fragments (occasionally unidentifiable) of agglutinated foraminifera as well.

Figure 2

The ¹⁴C dating of the core showed it to reflect the sediment structure extending as far back as to about 10 800 cal yrs BP (Fig. 3). The radiocarbon dates, albeit sparse, suggest continuous sedimentation and the absence of hiatuses. The depth–age model developed served as a timeframe for the foraminiferal record obtained from the core.

Figure 3

The deepest sediment layers (620–540 cm, c. 10 700–6900 cal yrs BP) were devoid of foraminifera which began to appear from the level of 535 cm up and were then present in most of the layers (except for 205–200, i.e. c. 3500 cal yrs BP, and 170–165 cm, i.e. c. 3000 cal yrs BP).

The study revealed the presence of a total of 13 foraminiferal taxa (including the collective Ammonia group; see below). The foraminifera found in the samples analyzed were represented by calcareous (hyaline) and agglutinated forms.

Most of the calcareous forms, belonging to the Rotaliidae, were assigned to the genus Ammonia and most likely represented the species A. beccarii and A. batava. However, the vast majority of the specimens occurred as organic shell linings (see above; Fig. 2) and decalcification prevented species identification. This, as well as the nomenclatural uncertainty regarding the taxonomic identification of individual Ammonia species (Hayward et al. 2004) and the lack of molecular confirmation of the identity of the specimens found, resulted in these forms being hereafter referred to as representing a higher-rank collective taxon, the Ammonia group. One layer (55–50 cm; c. 1400 cal yrs BP) showed the presence of representatives of the genus Elphidium (E. incertum, E. excavatum, and Elphidium sp.).

The agglutinated forms included Adercotryma sp., Ammoscalaria sp., Eggerelloides scaber, Miliammina fusca, Reophax cf. dentaliniformis, Textularia cf. kattegatensis and the Trochamminacea.

While some of the taxa identified (notably representatives of the Ammonia group) occurred throughout the foraminifera-containing part of the core (Fig. 3), others appeared either intermittently or were observed in some sediment layers only. In this context, it is interesting to note that Eggerelloides scaber, which first appeared in the 495–490 cm layer in the core (c. 7500 cal yrs BP), became a permanent member of the foraminiferal associations from the 200–195 cm layer (c. 3500 cal yrs BP) upward, and was virtually the only taxon found in the uppermost 45 cm layer.

The number of taxa was low throughout most of the core length, with basically monogeneric assemblages occurring at the opposite ends of the foraminifera-containing section. The first monogeneric assemblage occurred in the 540–500 cm section (c. 8100-7500 cal yrs BP) and consisted of the Ammonia group, while the other monogeneric (or even monospecific) assemblage, characterized by the exclusive presence of E. scaber, was recorded within the 40–5 cm layer (Fig. 3). The maximum taxonomic richness was observed within the 425–420 cm (five taxa) and 55–50 cm (six taxa) layers.

The abundance of foraminifera ranged from 0 to 1854.23 ind. g⁻¹ d.s. The abundance profile (Fig. 3) showed a sequence of peaks and troughs above the layer of sediment devoid of foraminifera, with three prominent peaks: the first (505.52 ind. g⁻¹ d.s.) within the 480–475 cm layer, the second being the maximum abundance observed within the 450–445 cm layer (c. 6900 cal yrs BP), and the third (761.05 ind. g⁻¹ d.s.) recorded within the 55–50 cm layer (c. 1400 cal yrs BP).

The proportion of calcareous to agglutinated foraminifera presents a very characteristic pattern (Fig. 3): calcareous forms accounted for up to 100% of the assemblage from the bottom of the foraminifera-containing section to the 190–185 cm layer. From this point on, the structure of the assemblage changes significantly. The proportions of calcareous to agglutinated forms begin to reverse, almost a mirror reflection-like, and the assemblage rapidly becomes dominated by agglutinated forms. Eventually, from about 50 cm in the core to the top, they account – almost without exception – for up to 100% of the assemblage, with E. scaber being the dominant (and frequently the only) taxon.

Based on the abundance of foraminifera along the core and other assemblage characteristics, the profile can be divided into five more or less distinct foraminifera-based units (F1–F5; Table 1) distributed from the bottom to the top of the core.

Unit F1 consists of the foraminifera-free part (up to 540 cm in the core; c. 8100 cal yrs BP). Unit F2 (540–470 cm; c. 8100–6900 cal yrs BP) features the first appearance of foraminifera and a gradual increase in their abundance to a peak (505.52 ind. g⁻¹ d.s.) within the 480–475 cm layer, followed by a decline. The foraminiferal assemblage in Unit F2 is dominated by the Ammonia group, with a slight admixture of agglutinated forms (notably Reophax cf. dentaliniformis, trochamminids and the sparse E. scaber). Unit F3, covering the 470–410 cm layer (c. 6900–6300 cal yrs BP), is the most distinct feature of the entire profile, with the maximum overall abundance (1854.23 ind. g⁻¹ d.s.) recorded at a depth of 450–445 cm in the sediment. The composition of the foraminiferal assemblage is most complex there, with the highest overall number of taxa recorded. Foraminifera of the Ammonia group dominate, but agglutinated forms (primarily R. cf. dentaliniformis, trochamminids, and Ammoscalaria sp.) are fairly well represented. Occasionally, the foraminiferal assemblage is accompanied by brackish-water ostracods (Cyprideis torosa). Unit F4, from 410 cm to 175 cm in the core (c. 6300–3000 cal yrs BP), is characterized by an abrupt decline in the foraminiferal abundance to generally low levels, with an undulating sequence of small peaks and troughs along the length of the core toward the top, and by a gradual shift in the community composition in which agglutinated forms (dominated by E. scaber) gain in importance from the 180–175 cm layer (c. 3000 cal yrs BP). Occasionally, the microfossil non-diatom assemblage contains valves of fresh- and brackish-water ostracods (Cypria ophtalmica and Cyprideis torosa) and cladoceran remains. Unit F5, from 175 cm (c. 3000 cal yrs BP) to the top of the core, is characterized by generally low foraminiferal abundances, including layers devoid of foraminifera, and by the domination of E. scaber. In addition, the unit shows (within 55–50 cm; c. 1400 cal yrs BP) an unusual peak in the foraminiferal abundance (761.05 ind. g⁻¹ d.s.), resulting from a strongly increased contribution of E. scaber, accompanied by the Ammonia group, Elphidium incertum and E. excavatum. Interestingly, all the Ammonia shells were decalcified, whereas those of Elphidium were not. The non-diatom microfossil assemblage found in the 55–50 cm layer contained also a collection of fresh- and brackish-water ostracods (Candona sp., Cyprideis torosa, Cytheropteron latissimum, Cytheromorpha fuscata, Loxoconcha elliptica, Acanthocythereis dunelmensis). This pronounced peak in the abundance of foraminifera is followed by a layer of very low abundances or a complete absence of foraminifera. The abundances increase again from 20–15 cm layer to the top (123.65–128.5 ind. g⁻¹ d.s. in the topmost 10 cm layer), with the assemblage being strongly dominated by E. scaber.

CCA produced four canonical dimensions, the first two of which proved statistically significant (p < 0.05). The ordination diagram of the first two canonical dimensions is shown in Fig. 4, while correlations of the canonical dimensions with the sedimentary parameters analyzed and the foraminifera groups are summarized in Table 2. The first canonical dimension accounts for 82% of the association between the sediment parameters and the foraminiferal abundance, and the first two canonical dimensions together explain 92% of the relationship. The first canonical dimension is positively associated with the content of magnesium (0.58) and biogenic silica (0.56); the second canonical dimension is most closely and positively associated with the mean grain size (0.87) and negatively associated with the biogenic silica content (−0.66). The first canonical dimension is positively associated with the Ammonia group (0.77), the Trochamminacea (0.55) and Reophax sp. (0.33), and negatively associated with E. scaber (−0.76). The second dimension is negatively associated with the Ammonia group (−0.61) and the Trochamminacea sp. (−0.55), no association with Reophax sp. (0.07) and E. scaber (0.05) being visible.

Figure 4

Analysis of Core 317980-3 from the Mecklenburg Bay (the south-western part of the Baltic Sea) showed the presence of a total of 13 foraminiferal taxa. Although generally indicative of low taxonomic diversity, the number of taxa found was more than twice as high as that reported by Binczewska et al. (2018), who found a total of six benthic foraminiferal species in their core collected from the Bornholm Basin (the central part of the Baltic Sea). This difference seems to be consistent with the general west-east diversity gradient of the Baltic Sea (Ojaveer et al. 2010; Snoeijs-Leijonmalm 2017).

Our study revealed remains of fairly abundant foraminiferal assemblages in the central MB sediment layers dating back to the onset of the Littorina stage in the area. Previous studies in MB (e.g. Hofmann & Winn 2000; Rößler 2006; Rößler et al. 2011) provided a rather inconsistent evidence regarding the presence of foraminifera in the area. Hofmann and Winn (2000), who collected two cores from MB (one in the innermost part, the Neustadt Bay, and the other in a more central location), found remains of foraminifera in the 0–310 cm section of their 494-cm long core from the Neustadt Bay, with the abundance of foraminifera forming peaks at 230 and 60 cm (the latter roughly coincident with our 55–50 cm peak). The abundances reported ranged within 2–51 ind. g⁻¹ wet sediment, a range not directly comparable with our abundance estimates, which were based on dry sediment. Hofmann and Winn (2000) provided no taxonomic identification, although they did mention that all the foraminifera found in their study represented the genera Ammonia and/or Elphidium, thus indirectly supporting our finding of the Ammonia group forms being dominant in the area (except for the upper part of our core). The other MB core of Hoffman and Winn (2000) yielded a single specimen, not identified taxonomically, found in the 38 cm sample.

Rößler (2006) referred to the abundance of foraminifera in her MB cores, although she did not provide any detailed list of the taxa found. She only mentioned that the foraminiferal assemblages appearing first in her MB cores were dominated by Ammonia batavus (an unaccepted synonym of A. batava; cf. World Register of Marine Species, http://www.marinespecies.org/aphia.php?p=taxdetails&id=112848 accessed on 6 December 2020), other conspicuously occurring species including A. beccarii, Cribroelphidium williamsoni (= Elphidium williamsoni in WORMS http://www.marinespecies.org/aphia.php?p=taxdetails&id=113291 accessed on 6 December 2020) and Haynesina germanica. In a subsequent analysis of four cores collected in MB, Rößler et al. (2011) did record the presence of foraminifera (see below), but again provided no information as to their taxonomic identity.

Based on published information and their own unpublished observations, Frenzel et al. (2005) produced a checklist of recent foraminifera from off the German Baltic coast, in which they summarized the knowledge about taxonomic richness in MB. They listed a total of 29 taxa (3 allogromiid, 14 agglutinated and 12 calcareous). Our material contains much fewer taxa, but all of them were already recorded from MB, even though they may have been referred to by Frenzel et al. (2005) under names that are not currently accepted (e.g. Nodulina dentaliniformis instead of Reophax dentaliniformis etc.). It is worth noting, however, that the list of MB foraminiferal taxa in Frenzel et al. (2005) does not contain Eggerelloides scaber, a species featuring so prominently in our material, particularly in the upper part of the core. Frenzel et al. (2005) did refer to it, but – quoting Murray (1991) – used the currently invalid names Eggerelloides scabrus and Egerella (sic!) scabra, the former being a dweller of higher-salinity areas in the open Baltic Sea and a key species of sand and mud associations, and the latter being used to denote a species (apparently the same one) included in the Ammotium cassis association occurring in the organic matter-enriched sediment in the inner Baltic Sea.

Among the calcareous taxa in our material, the most important were members of the Ammonia group, most likely Ammonia beccarii, a euryhaline species typical of shallow nearshore areas, estuaries and salt marshes, reported from numerous locations worldwide (e.g. Alve & Murray 1999).

The agglutinated foraminiferal taxa found in this work are characteristic of marginal marine areas (Murray & Alve 2011) and/or waters with reduced salinity. The most abundant agglutinated taxa in the samples included Eggerelloides scaber, Trochamminacea, Reophax cf. dentaliniformis, and Miliammina fusca. E. scaber has been described as “infaunal, detritivorous, subtidal, mainly shelf but tolerates salinity > 24 for most of the year and temperatures 1–20°C” (Murray & Alve 2011). Alve and Nagy (1986) showed E. scaber to be the most abundant species in Sandebukta on the west coast of Sweden, with no substrate preference. Christiansen et al. (1996) reported the species (as E. scabrus, currently an unaccepted synonym; World Register of Marine Species http://www.marinespecies.org/aphia.php?p=taxdetails&id=113939; accessed on 6 December 2020) as the most abundant foraminifera in the 130–140 mm layer of their short cores from the Kattegat. They described the species as tolerant of poor environmental conditions and resistant to pollution. Reophax dentaliniformis (referred to as Nodulina dentaliniformis by e.g. Murray & Alve 1999 and Frenzel et al. 2005) is a widely distributed species, found in both recent and sub-recent sediments of the Baltic Sea at different depths, as far east as the Gotland and Gdańsk Deeps (Hermelin 1983, 1987; A. Binczewska, pers. comm.). Miliammina fusca is another widely distributed species whose abundant occurrence was reported in the southern Baltic Sea (Brodniewicz 1965). It is very common and relatively abundant on the shallow (5–8 m) sandy bottom in the inshore areas of the Pomeranian Bay (southern Baltic Sea; T. Radziejewska, unpubl. data).

As in the study by Binczewska et al. (2018), the calcareous forms in our core (primarily the Ammonia group) occurred – in the vast majority of samples – as completely or partially decalcified organic test linings (cf. Fig. 2), somewhat similar to the photograph (Plate II 5) in Murray & Alve (1999). The post-mortem decalcification/dissolution of foraminiferal tests is a widespread and well-known effect (e.g. Christiansen et al. 1996; Leckie & Olson 2003; Buzas-Stephens 2005; Murray & Alve 2011; Haynert et al. 2012; Binczewska et al. 2018; Buzas-Stephens et al. 2018). To quote Murray and Alve (2011), “There is abundant evidence of syndepositional dissolution of calcareous tests in certain modern environments including both terrigenous and carbonate sediments” (p. 11). Christiansen et al. (1996) found abundant dissolution-affected foraminifera in the lower part (below 6 cm) of their core; they assigned most of those remains to Ammonia batava. Leckie and Olsen (2003) paid a particular attention to the taphonomic loss of calcareous species due to carbonate dissolution. They specifically referred to Ammonia, the preservation potential of which is adversely affected by acidic (low pH) conditions associated with organic-rich substrate. In Texas bays, Buzas-Stephens (2005) and Buzas-Stephens et al. (2018) found a high proportion of living foraminifers, almost exclusively Ammonia sp., with shells damaged by decalcification. They attributed this to acidic conditions prevailing in the sediment contaminated with heavy metals, and resulting in part from pyritization (evidenced by framboidal pyrite visible in shells of live specimens), a stress response in foraminifera. It can therefore be postulated, in line with conclusions of Binczewska et al. (2018), that the Ammonia-dominated foraminiferal assemblages observed in our core were developing in a brackish-water sedimentary environment, in fine-grained sediment substantially enriched with organic material (including that of plant origin, as exemplified by plant remains visible in the sieving residue examined), where – due to the decomposition of that material – the pore water pH was reduced (cf. Reaves, 1986).

The calcareous foraminifera test dissolution makes the sample processing technique very important. If samples contain decalcified tests, drying the sediment and examining dried samples for the presence of foraminifera, a widely adopted practice (e.g. Murray, 2002), would make it very difficult to spot the delicate shell linings of calcareous forms. We (and others: A. Binczewska and Z. Stachowska, pers. comm., Binczewska et al. 2018) experienced this difficulty before switching to examining sediment samples in water, whereby the linings were very well visible. As we used > 0.063 mm sieves, some fine sediment residue was lost during sample processing. However, we do not consider the loss of the finest material to be a factor that could affect the foraminiferal abundance data. More relevant in this respect could have been the fragmentation of test linings, as only fragments containing at least one whorl were counted, and whorl fragments were disregarded (see Sample processing above). For this reason, the total abundances presented may be treated as underestimates, but the extent of the underestimation cannot be assessed.

Interestingly, the calcareous forms found in MB cores by Rößler (2006) and Rößler et al. (2011) preserved their carbonate shells, enabling the authors to use those foraminifera for dating.

As opposed to the calcareous foraminifera, and in contrast to the findings of Binczewska et al. (2018), the agglutinated foraminifera found in our material were generally well-preserved, the numbers of damaged individuals and/or test fragments being generally very low.

We have grounds to assume that the reversal of the C/A ratio in the upper part of the core is not an artefact produced by dissolution of calcareous forms and the resultant “(…) higher abundance of agglutinated tests in the dead assemblages than could be accounted for by differential production between species.” (Murray and Alve, 2011). Because we analyzed non-dried sediment material, we would have recorded any organic test lining present in the sample, which was the case in the lower part of the core. We can therefore argue that the pattern observed reflects a true shift in the community composition. Assuming, after Levin (2003), that arenaceous tests are energetically less costly to form and maintain than the calcareous ones, the strong preponderance of agglutinated foraminifera in the upper part of the core can be taken as evidence of environmental conditions becoming increasingly stressful in terms of e.g. reduced oxygenation of the near-bottom water layer and pore water and/or reduced salinity, for calcareous forms to exist (e.g. Iglikowska & Pawłowska 2015). Leckie and Olson (2003) found a marked increase in the abundance of calcareous taxa to accompany the transition from brackish marginal habitats to open neritic conditions. Similarly, Valchev (2003) attributed a shift in the C/A ratio in favour of arenaceous (i.e. agglutinated) foraminifera to low-salinity water undersaturated in calcium carbonate. We may thus conclude that a marked change in the C/A ratio observed in our core signified a transition to a less saline and more stressful environment in MB prevailing to date.

No foraminifera remains were found in the deepest part of the core (Unit F1; 615–610 and 615–540 cm), i.e. the sediment older than 10 700 to 8 100 cal yrs BP (cf. Fig. 3), attributed to the final stage of the Baltic Ice Lake (620–615 cm) and especially the Ancylus Lake phases (Kostecki et al. 2015). On the other hand, the sediment throughout the 615–540 cm layer contained extremely abundant cladoceran remains (carapace fragments and ephippia). Although no formal study on cladoceran proxies was carried out, the remains were identified as representing carapace fragments of the dominant Bosmina longispina, accompanied by Alona quadrangularis, A. affinis, Chydorus sphaericus, Ch. piger, Leydigia leydigi, Pleuroxus uncinatus, Alonella nana, Bosmina longirostris, Eurycercus lamellatus, Acroperus harpae, Camptocercus rectirostris and L. acathocecoides (A. Kaniak, pers. comm.). These species, representing a typically freshwater assemblage, testify to the lacustrine nature of the environment recorded by the sediment in this layer, the presence of some cold-water species (e.g. B. longispina, Ch. sphaericus, A. nana; Zawisza & Szeroczyńska 2007) evidencing the low-temperature phase in the MB evolution. This oldest, Ancylus Lake-derived section of the core corresponds to the Ancylus Lake (AL) section of the Bornholm Basin core analyzed by Binczewska et al. (2018), also lacking (or nearly lacking) foraminifera and featuring abundant cladoceran remains. According to Binczewska et al. (2018), however, their AL section corresponded to the period of 7100–6900 cal yrs BP, thus later than the Ancylus Lake stage occurring in MB. It seems that while the Ancylus Lake stage was still prevalent in the Bornholm Basin, MB in the western part of the present-day Baltic Sea began to experience a shift toward the next stage of the Baltic Sea evolution – the Littorina Sea stage.

The overlying layer (Unit F2, 540–470 cm; c. 8100–6900 cal yrs BP), where foraminifera begin to appear, marks the onset of the transition in the nature of the area toward the Littorina Sea stage. The sporadic hydrobiid snail shell fragments (cf. Table 1) may be indicative of the shallowness of the area. The subsequent layer (Unit F3, 470–410 cm; c. 6900–6300 cal yrs BP) coincides with the beginning of the early-and mid-Littorina Sea stage (LS) of Binczewska et al. (2018). In our core, this stage is characterized by the highest abundance of foraminifera (cf. Fig. 3), also reported by Binczewska et al. (2018), and the assemblage is dominated by the calcareous Ammonia group (Elphidium spp. in the case of the Bornholm Basin). The sediment characteristics (cf. Table 1 and Kostecki et al. 2015) are indicative of a deep low-energy depositional environment. The 410-175 cm layer (Unit F4; c. 6300–3000 cal yrs BP, coinciding with the later part of the early- and mid-LS stage of Binczewska et al. 2018) shows a decline in the abundance of foraminifera and a change in the assemblage structure, whereby the previously dominant calcareous forms give way to the domination of agglutinating foraminifera (E. scaber), which is particularly well visible in the uppermost unit (F5; c. 3000 cal yrs BP-present; cf. Table 1). As already mentioned, agglutinated foraminifera are known to usually dominate at lower salinities, and a change from a calcareous to an agglutinated assemblage visible in the sedimentary record may be taken as evidence of a change (reduction) in salinity (Leckie and Olson, 2003; Buzas-Stephens et al. 2018). The stage in question corresponds to the late-LS stage of Binczewska et al. (2018), characterized by low salinities and the domination of Reophax dentaliniformis, an agglutinated species occurring with low densities, and the scarcity or absence of calcareous foraminifera.

The pattern emerging from the analysis of our core is confirmed by the CCA results (cf. Fig. 4). The first CCA dimension (with the highest and significant correlation with magnesium and biogenic silica contents) can be considered to be related to salinity fluctuations, with magnesium regarded as a salinity proxy (cf. Ruiz-Agudo et al. 2010). Elevated salinity can be inferred from the higher contents of magnesium and biogenic silica that accompanied the abundant presence of the Ammonia group foraminifera (and – to a much lesser extent – the Trochamminacea). The biogenic silica content reflected the presence of an abundant marine diatom flora (Kostecki et al. 2015), also indirectly indicating an environment with higher salinity. Rößler (2006) observed the MB calcareous foraminifera peak to coincide with peaks in the sediment calcium carbonate, with both magnesium and calcium being components of biogenic marine carbonates (Ruiz-Agudo et al. 2010). The Ammonia group (and to a lesser extent Reophax sp. and the Trochamminacea) were the most abundant foraminifera during the period of increasing salinity, while the presence of E. scaber should be related to the period of decreasing salinity. In this context, it is noteworthy that Polovodova et al. (2009) considered E. scaber as an indicator of increased salinity: they interpreted the decline in the species’ abundance in the Flensburg Fjord as an indication of reduced incidence of saline oxygen-rich water inflows. In our study, however, E. scaber was an absolute dominant in the uppermost stratigraphic unit related to the period of decreasing salinity in the Baltic Sea.

The second canonical dimension was strongly influenced by the mean grain size, and thus can be inferred to reflect changes in the depositional energy status of the area and its water depth. Higher values of the second dimension are related to shallow-water dynamic sedimentation affected by waves and currents, while low values relate to calmer sedimentation conditions at deeper water favoured by foraminiferal assemblages featuring the Ammonia group. The low correlations of trochamminids and Reophax sp. with the second dimension confirmed their preference for shallow-water habitats (Frenzel et al. 2005).

The foraminifera-based stratigraphy inferred from the analysis of Core 317980-3 (Table 1) is broadly similar to that based on the diatomological analysis of the same core (Kostecki et al., 2015). There are, however, some differences. The diatom LDAZ1 and LDAZ2 considered together coincide with the foraminifera-based unit F1 covering the core section devoid of foraminifera. The remains of freshwater diatoms in the fully lacustrine Ancylus phase represented by this section allowed us to fine-tune the resolution of various periods during that stage. On the other hand, the remaining core section reflects two diatom-based units (LDAZ3 and LDAZ4), whereas the foraminiferal analysis identified four units (F2–F5), refining the division resulting from the diatom analysis and coinciding broadly with the stratigraphy based on geochemical variables analyzed by Kostecki et al. (2015).

In view of the results presented in this study, we can conclude that the analysis of foraminifera complements the set of palaeoproxies used previously and provides additional information on the Holocene evolution of MB in its post-Ancylus Lake stages.