Regional background

The location of the BSB in a glaciation-sensitive high northern latitude has resulted in a highly dynamic development during its recent geological history. Areas of the BSB have been repeatedly covered by recurring Quaternary glaciations over northern Europe, with subsequent deglaciations resulting in largely differential uplift across the region of the BSB and its drainage area. The last interglacial–glacial cycle is an excellent analog of the various processes that the basin has been exposed to during previous glacial cycles, with geological deposits in the Baltic Basin and the surrounding region acting as archives of its history.

The Eemian interglacial (~130–115 k.y. before present [BP]) is partly documented in the North Greenland Ice Core Project (NGRIP) ice core (North Greenland Ice Core Project Members, 2004) and in low-resolution deep-sea records (land-sea correlation, Sánshez-Goñi et al., 1999). However, detailed Eemian marine shelf records from northern Europe are very scarce and generally contain only fragmented paleoenvironmental information. The same is true of the early Weichselian interstadials (marine isotope Stage [MIS] 5c and MIS 5a).

The buildup to the Last Glacial Maximum (LGM), MIS 3 (~60 to 25 k.y. BP), is characterized by many rapid switches between short cold and longer warm periods. The latter are known as Dansgaard-Oeschger (D/O) events, which are evident in both the Greenland ice cores and deep sea records. The D/O events are part of the “Bond cycles,” each one ending with a Heinrich event disrupting North Atlantic thermohaline circulation and causing cooling in the Northern Hemisphere (e.g., Dickson et al., 2008; Grimm et al., 2006; Bond et al., 2001). The complexity of these cycles is intricate: whereas warming can destabilize ice sheets and produce surges and icebergs, a cooler climate increases sea ice cover and consequently increases Earth’s albedo and promotes further ice sheet growth. However, the feedback from increased sea ice cover in the North Atlantic on the SIS, and hence the climatic effects on northwestern Europe, is presently poorly understood.

The long-term paleoenvironmental development since the LGM at ~20 k.y. BP (MIS 2) is fairly well understood, and the drastic climatic shifts during the Last Termination as recorded in the Greenland oxygen isotopic record are mirrored in the Baltic varved glacial clay record (Andrén et al., 1999, 2002).

The Late Pleistocene and Holocene climatic history of Central and Northern Europe is closely connected to the large-scale meteorological processes of the wider North Atlantic region, which also controlled the climate development of the North American continent. Different terminologies are used for the glacial history of Europe and North America. Table T1 provides a correlation of terminologies used for the two regions and the approximate ages and marine isotope stages.

Geological history of the Baltic Sea Basin during the last glacial cycle

The BSB has undergone many glaciations during the Quaternary. During the Last Interglacial (the Eemian), the BSB was a larger and more saline sea than the present Baltic Sea (Funder et al., 2002). A thin and incomplete marine Eemian sediment sequence has been found as a thrusted sequence in a coastal cliff in the western Baltic (e.g., Kristensen and Knudsen, 2006; Gibbard and Glaister, 2006), but no complete sequence of sediments deposited during the Eemian has been recovered in situ in any part of the Baltic Sea. Consequently, we have incomplete knowledge about the BSB during the Eemian interglacial. Two subsequent early Weichselian glacial advances (MIS 5d and MIS 5b) may have only reached as far south as ~60.5°N, leaving the central and southern BSB free of ice. Glacial events during MIS 4 were recorded in sediments from (north)western Finland at ~64°N (Salonen et al., 2007), and the first Baltic ice lobe advance into Denmark is dated to 50–55 k.y. BP (Houmark-Nielsen, 2007; Larsen et al., 2009). Based on these records, an isolated lake probably covered large parts of the BSB until at least 60 k.y. BP. This portion of the BSB acted as a depocenter between 55 and 60 k.y. BP, and accumulated sediments most likely hold a high-resolution record of climate-related changes across the large catchment area of the BSB.

The complex Danish glacial stratigraphy and several 14C dates (26–35 14C k.y. BP; Andrén and Björck, unpubl. data) of peat and lake (35 and 65 m below present sea level, respectively) sediments covered and underlain by till and clays in the southern BSB indicate a highly oscillating ice margin. These data show that the southern BSB was partly free of ice on several occasions during MIS 3 prior to the LGM at ~22 k.y. BP when the Weichselian ice sheet covered the entire BSB (Houmark-Nielsen and Kjær, 2003; Larsen et al., 2009). It also implies that the central parts of the Hanö and Bornholm Basins were lacustrine systems during ice-free conditions and that glacial erosion, at least locally, was insignificant, leaving much of the MIS 3 and older archives intact. Furthermore, glacial erosion must have been minimal in the deepest basins during these inferred surge-like events. This part of the BSB may therefore hold a record of Heinrich and D/O events and answer questions about what if any role the SIS played in these enigmatic oscillations.

Deglaciation and early postglacial stages

The deglaciation of the southern BSB between 22 and 16 k.y. BP was highly complex, with major deglacial phases interrupted by intriguing stillstands and re-advances (Houmark-Nielsen and Kjær, 2003; Larsen et al., 2009) (Table T2). The Baltic Ice Lake (BIL), dammed in front of the retreating ice front (Fig. F2), released large amounts of freshwater into the North Atlantic during the early stage of the deglaciation between ~16 and 11.7 k.y. BP. This occurred both through the Öresund Strait and through south central Sweden (Gyldenholm et al., 1993; Björck, 1995; Majoran and Nordberg, 1997; Jiang et al., 1998). During the final drainage of the BIL at ~11.7 k.y. BP, almost 8000 km3 of freshwater was released rapidly into the North Atlantic (Jakobsson et al., 2007), but the effects on North Atlantic thermohaline circulation may have been only minor (Andrén et al., 2002) (Fig. F3). However, it has also been suggested that this sudden freshwater discharge may have triggered the cold Preboreal oscillation (Björck et al., 1996) and ice advances in northern Norway (Hald and Hagen, 1998).

The next Baltic Sea stage, the Yoldia Sea (YS), coincides with the onset of the Holocene epoch (Walker et al., 2009) and the associated rapid warming. In fact, the thicknesses of glacial varves in the northwestern Baltic proper and δ18O values in the Greenland Ice Core Project (GRIP) ice core display a noticeably similar pattern over a 150 y Younger Dryas to Pre-boreal transition period (Andrén et al., 1999, 2002). These records show a distinct increase in sedimentation rate as the ice sheet began to melt and rapidly retreat (Fig. F3). The following few hundred years were characterized by rapid deglaciation of the SIS. Relative sea level changes of the YS played an important role and were the result of a combination of rapid regression in the recently deglaciated regions and normal regression rates in southern Sweden (1.5–2 m/100 y).

Despite the fact that the passage westward to the North Sea was open and the YS was at the same topographic level as the North Sea, it took ~300 y before saline water could enter through the fairly narrow straits of the south-central Swedish lowland (Fig. F4). This brackish phase has been documented by varve lithology, geochemistry, and marine/brackish fossils, such as Portlandia (Yoldia) arctica. Occasionally, this phase is represented by sulfide banding, implying a halocline. The maximum duration of this brackish phase has been estimated to 350 y (Andrén et al., 2007), although some records indicate that it only lasted 70–120 y (Andrén and Sohlenius, 1995; Wastegård et al., 1995). Because of the high isostatic rebound rate in south-central Sweden, the strait rapidly shallowed, which together with the large outflow of meltwater from the retreating ice sheet prevented further inflow of saline water to the Baltic (Fig. F4). This turned the YS into a freshwater basin again, although there was still an open connection to the North Sea in the west. At the end of this stage, the ice sheet had receded to the north and most of today’s BSB was deglaciated, with the exception of the Bothnian Bay.

As the outlets west of Lake Vänern became shallower, subsequent damming forced the water level inside the Baltic Basin to rise, and the next stage, the Ancylus Lake (AL), began. The sediments of this large freshwater lake generally contain little organic material, which may be explained partly as a result of the meltwater inflow to the Baltic from the final deglaciation of the SIS and is also due to the young soils of the recently deglaciated drainage area. Together, this created an aquatic environment with low nutrient input and hence low productivity. This environment did not promote the formation of a halocline but led to a well-mixed oxygenated water body. The relatively common sulfide-banded sediment of this stage can probably be explained by H2S diffusion from younger organic-rich sediments (Sohlenius et al., 2001).

Areas south of Stockholm–Helsinki that had experienced shoreline regression since the deglaciation now entered a transgressive phase, marking the onset of the AL stage. This stage is documented by several shore displacement curves from this area but also by simultaneous flooding of pine forests growing along the coasts of the southern Baltic Basin. In the southernmost Baltic, a transgression is recorded by submerged pine trees and peat deposits dated between 11.0 and 10.5 k.y. BP (Andrén et al., 2007).

Whereas areas to the north experienced a slow regression, the extent of the transgression in the south varied depending on whether areas were isostatically rising or submerging (Fig. F5). A large number of 14C dates from peat as well as tree remains (mainly pine) from the beaches suggest the time span of this so-called Ancylus transgression can be estimated to ~500 y. The isobases over southern Sweden show that the water level of the Baltic was higher than the sea in the west, indicating that the AL was dammed. The total impact of this damming is estimated to have raised the Baltic ~10 m above the level of the North Sea (Björck et al., 2008), meaning that the isostatically submerging areas in the southernmost Baltic experienced a transgression larger than that.

The transgression and flooding in the south, as a result of the differences in isostatic rebound between the northern and southern Baltic Basin, would have resulted in a new outlet in the south. Because the outlet through Öresund had been uplifted more than other possible outlets farther south, it is likely that the new outlet was located somewhere in these southern areas (Fig. F1). Available data indicate that the Darss Sill area, between Darss and Møn, was inundated by the AL. The idea of a sudden and large drainage of the AL through the so-called Dana River flowing through Mecklenburg Bay and Fehmarn Belt, along the east side of Langeland and out through the Great Belt to Kattegat as proposed by von Post (1929) was proven impossible by Lemke et al. (2001). However, initial erosion of soft Quaternary deposits from the riverbed might have opened a connection and caused an initial lowering of the AL of ~5 m, followed by a period of rising sea level. It has been proposed that this resulted in a complex river system through Denmark comprising river channels and lakes (Bennike et al., 2004), with a steadily lower fall-gradient as the sea level rose. As the sea level in Kattegat had reached the level of the AL inside the Baltic Basin, it was possible for saltwater to penetrate all the way through the long river system into the Baltic. The first weak inflows of saline water are recorded simultaneously in the Bornholm Basin (Andrén et al., 2000b) and the Blekinge archipelago (Berglund et al., 2005) at ~9.8 k.y. BP. It is therefore possible to define the end of the Ancylus Lake as being concurrent with the time at which the AL level was at the same approximate level as the North Sea and the first signs of any marine influence since the YS are recorded. This period of low and fluctuating saline influence lasted longer than a millennium and has been named the Initial Littorina Sea (Andrén et al. 2000b).

The Littorina Sea

The onset of the next stage, the Littorina Sea (LS), can often be recognized as a marked lithologic change in Baltic Sea cores. The onset may be represented by a distinct increase in organic content and an increasing abundance of brackish marine diatoms (e.g., Sohlenius et al., 2001). Whether this sudden increase in organic carbon content is exclusively coupled to changes in primary production or if it is partly due to a better preservation of organic carbon during anoxic conditions has been a matter of discussion (Sohlenius et al., 1996). Flocculation of clay particles and subsequent rapid sedimentation resulting in an increase in the Secchi depth has also been proposed as a contributor to an increase in primary production (Winterhalter, 1992). The distribution of trace elements in sediments, especially the enrichment of barium and vanadium, which is linked to cycling of organic carbon, implies that increased productivity in the basin is a plausible cause of the rise in organic carbon content (Sternbeck et al., 2000).

The global melting of the large ice sheets over a couple of millennia caused a 30 m rise in absolute sea level (Lambeck and Chappell, 2001). A consequence of this is the flooding of the Öresund Strait, believed to be the main mechanism behind the onset of the LS. The outlets/inlets at Öresund and Great Belt widened and became deeper, resulting in greater flow of increasingly saline water (Fig. F6). Maximum postglacial salinity was reached at ~6 k.y. BP (e.g., Westman and Sohlenius, 1999). Gustafsson and Westman (2002) proposed, based on model calculations, that the salinity variations during the last ~8 k.y. BP are only partly explained by changes in the morphology and depths of the sills in the inlet area. They suggest that a major cause of the salinity changes may have been variation in the freshwater input to the basin. It was demonstrated that the freshwater supply to the basin may have been 15%–60% lower than at present during the phase of maximum salinity around 6 k.y. BP. Climate-driven long-term freshwater discharge variability may also have been an important factor controlling salinity and stratification, hence the distribution of hypoxia in the Baltic Sea during the last ~8 k.y. (Zillén et al., 2008). Periods of deepwater hypoxia in the open Baltic Basin are evident in the sediment record as extended sections of laminated sediment (Sohlenius and Westman, 1998; Zillén et al., 2008). These are considered a direct result of this salinity stratification together with increased primary production. Increased upward transport of nutrients from the anoxic bottom water has been suggested as an explanation for the enhanced primary productivity at the Ancylus–Littorina transition (Sohlenius et al., 1996). The LS experienced a sustained period of hypoxia between 8 and 4 k.y. BP (Zillén et al., 2008). The inflowing marine water in the Baltic Sea probably caused the release of phosphorus from the sediments, thereby enhancing the growth of nitrogen-fixing cyanobacteria (Bianchi et al., 2000; Borgendahl and Westman, 2007; Kunzendorf et al., 2001).

Because of dating uncertainties of reworked “old” carbon in the sediment, it has not been possible to absolutely date the transition from fresh to brackish water. In general, 14C dates between 8.5 and 8 k.y. BP are very common for the onset of this shift in salinity and sedimentary regime (Sohlenius et al., 1996; Sohlenius and Westman, 1998; Andrén et al., 2000a). However, an optically stimulated luminescence (OSL)-based age-depth model presented by Kortekas et al. (2007) suggests an age of 6.5 k.y. BP for the same shift in the Arkona Basin. 14C ages of both bulk sediment and bivalves from the same core give ages older than the OSL ages yet younger than the expected 14C age of 8.5–8 k.y. BP. As diatom analyses were not carried out by Kortekas et al. (2007), it may be that the shift dated was not the same as that determined in the other studies.

A climatic event at 8.2 k.y. BP has been recognized in both marine and terrestrial records from northern Europe (e.g., Alley et al., 1997; Klitgaard-Kristensen et al., 1998; Tinner and Lotter, 2001; Snowball et al., 2002). A drift gyttja layer dated to 8.2–8.1 k.y. BP from coastal sites in the Blekinge area in the southern BSB has been interpreted as representative of a regional catastrophic event, possibly the result of disturbance in the climate regime in the North Atlantic region (Berglund et al., 2005). In the central Bornholm Basin, a similar stratigraphic unit is recorded and dated to ~8.0 k.y. BP. As it is associated with a hiatus, the dating is uncertain (Andrén et al., 2000b; Andrén and Andrén, 2001), but it may be related to the drift gyttja layer in the Blekinge area and the associated catastrophic event. The full consequence of the short climatic cooling during the 8.2–8.1 k.y. BP event in the BSB is presently not well understood.

The present Baltic Sea

Around 6–5 k.y. BP, the eustatic sea level rise gradually ceased and the remaining slow rebound resulted in shallowing of the Öresund and Great Belt Straits and decreased salinities in the BSB. There have been several attempts to estimate the paleosalinity of the Baltic Basin, for example that of Gustafsson and Westman (2002). They used the presence or absence of mollusks and a silicoflagellate species to infer different salinity intervals of the last 8 k.y. Using stable carbon isotopes, Emeis et al. (2003) reconstructed Baltic salinity fluctuations throughout the Holocene. However, a comparison of these two salinity reconstructions (Zillén et al., 2008) shows large discrepancies. The reconstruction by Gustafsson and Westman (2002) shows decreasing salinity following a maximum value between 5 and 6 k.y. BP, whereas the study by Emeis et al. (2003) suggests increasing salinity during the last 2 k.y. Using the strontium isotopic ratio in carbonate mollusk shells, Widerlund and Andersson (2006) inferred paleosalinity and quantified salinity with a precision greater than ±5%. A drawback with the method, however, is that it can only be used when carbonate shells are preserved in the sediments. Donner et al. (1999) suggest salinity ~4‰ higher than present in the coastal areas of the Gulf of Finland and the Bothnian Sea between ~7.5 and 4.5 k.y. BP based on the 18O/16O ratio in mollusk shells.

After ~4 k.y. BP, salinity decreased and oxygen conditions improved considerably in the bottom waters (Gustafsson and Westman, 2002). This coincided with the onset of the neoglaciation in northern Europe with a more humid and cold climate (Snowball et al., 2004). Such a decrease in salinity may be the result of a shift to colder and wetter conditions, which probably increased the net precipitation in the watershed, leading to increased freshwater supplies to the basin (Gustafsson and Westman, 2002). This scenario, in combination with increased wind stress over the Baltic Sea, would lead to more efficient exchange of oxygen across the halocline because of a weakened halocline and enhanced vertical mixing (Conley et al., 2002). This scenario would also explain the reduction of the hypoxic zone around 4 k.y. BP as a result of more oxic bottom water conditions (Zillén et al., 2008).

The causes for widespread hypoxia during the middle-late Littorina Sea are not fully understood, but the fact that it occurred simultaneously with the growing population around the Baltic Basin may indicate forcing via eutrophication as a result of increased anthropogenic influence. Population growth and large-scale changes in land use that occurred in the Baltic Sea watershed during the Medieval Period between AD 750 and 1300 has been suggested as a triggering mechanism behind the expansion of hypoxia at that time (Zillén et al., 2008; Zillén and Conley, 2010). Alternative hypotheses suggest that the development of hypoxia in the open Baltic Sea over the past 1000 y is mainly driven by the climate system (e.g., Kabel et al., 2012). It is therefore likely that the late record of hypoxia in the Baltic Sea may be a result of multiple stressors, the interaction of climatic and anthropogenic influences.

Around the turn of the last century, hypoxia again appeared in the Baltic Sea recorded as laminated sediments deeper than 150 m in the Eastern Gotland Basin (Hille et al., 2006). This period corresponds to the onset of the Industrial Revolution when the European population size increased rapidly (about six-fold since AD 1800) and great technological advances took place in agriculture and forestry (Zillén and Conley, 2010). It also correlates with climate warming that has lasted most of the twentieth century. The present-day eutrophication (e.g., Elmgren, 2001) may be partly explained by the ongoing climate warming (Andrén et al., 2000a; Leipe et al., 2008). Superimposed on this climate warming are the effects of increased nutrient discharge related to a growing population and the intensive use of synthetic fertilizers on arable land after World War II (Elmgren, 1989). It is a major scientific challenge for the future to understand and predict the interactions between climate and anthropogenic forcing on the Baltic Sea ecosystem.

Microbial life in the Baltic seabed

The Baltic seabed holds a record of past biota of plants and animals that lived in the BSB ecosystem or in the surrounding catchment area. The fossil remains are evidence of strong changes in the living communities that resulted from shifts in climate and environment during the glacial–interglacial history. The seabed also harbors extant communities of living microorganisms that thrive on fossil organic matter that is now buried deeply beneath the seafloor.

Such microorganisms have been found abundantly tens to hundreds of meters deep in the sediments of the worlds’ oceans with total numbers that equal those of all microorganisms in the water column (Kallmeyer et al., 2012) (Fig. F7). These microorganisms play a key role in the slow degradation of organic matter and remineralization of nutrients. With increasing depth and age of the deposits, the remaining organic matter becomes increasingly recalcitrant and the availability of nutrients and energy for the microorganisms becomes very low. The combination of large microbial communities and very low carbon flux means that the individual cells gain little energy and therefore have correspondingly slow growth and long generation times of hundreds to thousands of years (D’Hondt et al., 2004; Parkes et al., 2005; Biddle et al., 2006; Hoehler and Jørgensen, 2013).

Many fundamental questions remain before we understand the life of these microorganisms in the deep subseafloor biosphere. It is not understood whether the deep biosphere is inhabited by a unique community of organisms that are specifically adapted to this mode of life under apparent long-term starvation. It is also not understood whether the diversity of microorganisms from past sediment surfaces still dominates the microbial biome after they have been buried for thousands of years, thus remaining as a “paleome” from former times (Inagaki et al., 2005).

The BSB provides unique possibilities to address these and many other questions. In contrast to most other sediments cored during IODP, the surface environment of the BSB has changed dramatically over its glacial–interglacial history so that the abundance and diversity of the buried microbial communities can now be related to both the present and the past environment.