Factors controlling the formation of laminated sediments in Lake Kierskie
During the monitoring period, physical and chemical characteristics of Lake Kierskie waters were interrelated with the primary productivity in the epilimnion, sediment production and the annual cycle of water mixing and stratification.
The late autumn
The sediment total mass flux to the sediment traps between 13th November and 18th December 2015 constituted more than 17% of the total mass of the sediment accumulated during the monitoring period (Table 1), and thus was one of the highest during the monitoring period (Fig. 6a). The biomass of phytoplankton measured on 13th November and 18th December was very low, sharply lower compared to further seasons of the monitoring study (Fig. 5; ESM 2), which is also reflected in the lowest chlorophyll-a concentrations and the best water transparency in the whole study period (Fig. 6i, j). Therefore, material accumulated in the traps during the late autumn cannot be explained by the sediment flux from the epilimnion related to phytoplankton activity. High CaCO3 contribution to the sediment (approximately 60% of the total sediment flux; Table 1; Fig. 6b) contradicts the negative calcite SI values (between − 0.67 and − 1 at all depths; Fig. 6h), which indicate suitable conditions for calcite dissolution. Therefore, we suggest that sedimentation during the late autumn must have been controlled by the processes related to water turnover in the lake, the latter indicated by the vertically unified phytoplankton biomass and physico-chemical parameters of the water (ESM 1; Fig. 4). In Lake Kierskie the processes of sediment resuspension activated by water movement during the lake turnover, and subsequent redeposition of the material are regarded as critical for the increased sediment flux during the late autumn. Sediment resuspension and redeposition are well known from lakes (Charlton and Lean 1987) and have been shown to result in a secondary flux of resettling material in sediment traps (Bloesch 1994; Evans 1994). The redeposited sediment is expected to form distinct laminae within the yearly sediment record in Lake Kierskie, however, this is still to be proven by microscopic study of laminated sediments. Preservation of the laminated sediments in the deepest part of the lake studied indicates that the resuspended material must have originated mostly from the shallower parts of the lake (Tylmann et al. 2012). Resuspension of the sediments from the deepest part of the lake was either absent or small enough to allow preservation of the seasonal flux in the sediments.
At the breakdown of water stratification in a lake an additional sediment flux may origin from the release of sediment particles trapped in the thermocline/chemocline, which then settle to the lake bottom and contribute to the sediment record (Punning et al. 2003; Rull et al. 2017). During the summer of 2016, the vertical profiles of both water temperature and conductivity in Lake Kierskie showed distinct shifts towards lower and higher values, respectively, at the depths of 6–10 m (Fig. 4). Such thermal and chemical gradients result in water density changes where some of the particles settle and can be released only after the gradients diminish or disappear as the water cools in the autumn and turnover of the water column establishes.
Focusing of the suspended material during the water turnover is indicated by a nearly double mass of the sediment accumulated in the sediment traps at 30 m, compared to the shallower traps (Fig. 6a; Table 1). Therefore, the actual sediment mass flux to the lake bottom can differ laterally and is expected to be greatest in the deepest part of the lake. Sediment focusing with increasing water depth was observed in other lakes as well (Bloesch 1994; Evans 1994; Leemann and Niessen 1994). In Lake Stechlin (Mothes 1985) the total mass flux of the sediment increased by a factor of 1.7 between 20 and 40 m of the water depth.
The mean sediment flux to the traps was the lowest between 18th December 2015 and 17th March 2016 (Fig. 6a) when primary productivity contribution to the sediment accumulated was significantly limited. Low sediment flux points to the limited role of resuspension and redeposition during the winter stratification, however, those processes cannot be totally excluded. Slow sedimentation of the smallest particles from the suspension is regarded to contribute to the winter sediment flux in Lake Kierskie. The greater total sediment flux in the deeper sediment trap (Table 1) indicates sediment focusing with increasing depth. The proportions between CaCO3, OM and the remaining material (likely SiO2) in the sediments collected during the winter differ from the late autumn, spring and summer seasons (Table 1). This change relates mainly to the decreased share of CaCO3 and an increased amount of the remaining material. A smaller percentage of CaCO3 in the sediments is a consequence of lack of calcite precipitation in the epilimnion and the dissolution of the carbonates indicated by the negative calcite SI values (Fig. 6h; ESM 1).
Spring and summer
In contrast to the substantial contribution of redeposited material to the sediment traps in the autumn, we did not observe increased sediment accumulation in the traps during the water turnover in the early spring (Figs. 4, 6a). The difference observed may relate to the length of the mixing period. The prolonged period of relatively equal air temperatures during most of October, November and December 2015 (Fig. 6e) was suitable for establishing of the equalized temperatures of lake waters and long water mixing (Fig. 4). On the contrary, in the early spring temperatures increased fast and the length of water mixing was limited (Figs. 4, 6e). It seems possible that thickness of the laminae deposited during the water turnover is a potential indicator of the duration of the water mixing, with thicker laminae formed in the years with longer homothermy in the water column and long period of water mixing. However, this suggestion should be verified by further studies.
Phytoplankton bloom started already during the spring water mixing when nutrients became available in the epilimnion. The phytoplankton assemblage was dominated by centric diatoms Stephanodiscus hantzschii and Cyclotella sp. (Figs 4, 5; ESM 2). Some of the diatom species could have bloomed under the ice (Vehmaa and Salonen 2009) that discontinuously occurred on the lake. The coincidence of the maximum biomass and abundance of the diatom species (Fig. 5) with the water mixing during the lake turnover (Fig. 4) in the early spring is typical for this group of phytoplankton (Reynolds 2006). Regarding their build, i.e. lack of flagella and therefore the disability to move, and heavy siliceous shell, diatoms are a group of phytoplankton characterized by the fastest sedimentation rates among the phytoplankton (Sommer 1984; Padisák et al. 2003). As was evidenced by Smetacek (1985), diatoms of the size of S. hantzschii settle down very rapidly and within one day they can reach the sediment traps installed at the depth of 30 m. What keeps diatoms in suspension is the presence of water movement (waves, mixing; Fuchs et al. 2016). Under windless conditions and establishing of the stratification, the sedimentation rate of diatoms increases instantly (Padisák et al. 2003). Accordingly, in Lake Kierskie mixing of the water column in March (Fig. 4), allowed to prolong the time of diatom suspension (Fig. 5; ESM 2) and contributed to the bloom of this group of phytoplankton. Interestingly, the maximum biomass and abundance of the diatom species preceding the onset of stratification in Lake Kierskie (Figs. 4, 5; ESM 2) is in contrast to observations by Kienel et al. (2017) from Tiefer See, who evidenced the peak of phytoplankton concentrations, dominated by diatom assemblages, right after the onset of stratification, and concentration of the diatoms within the photic layer of the lake water. The discrepancy observed may result from the different methodology used. The diatom data of Kienel et al. (2017) are based on sediment samples collected from traps in 15-day intervals, therefore with a time lag, whereas we have described an actual phytoplankton assemblage and corresponding limnological conditions.
In Lake Kierskie the maxima of the phytoplankton biomass are clearly reflected in the physical and chemical properties of water, namely in the second-highest content of chlorophyll-a, during the monitoring period, the peak of O2 concentration in water, and marked decrease in Secchi disk visibility (Fig. 6i, j). Domination of the phytoplankton assemblage by centric diatoms (Fig. 5; ESM 2), resulted in a very strong increase in SiO2 content in the sediments, considerably higher to the sediment traps installed at 30 m (Fig. 6d), indicating sediment focusing. Observations from Lake Kierskie agree with the increased sedimentation rate observed after spring diatom bloom in lakes Constance (Grossart and Simon 1998) and Holzmaar (Moschen et al. 2006).
The spring phytoplankton bloom is also recorded in increased OM content in the sediments (Fig. 6c). Intensive primary productivity in the pelagial zones of lakes with waters rich in dissolved calcium and bicarbonates results in precipitation of CaCO3 (Rodrigo et al. 1993). The biological precipitation of calcium carbonate occurs when the CO2 assimilation during photosynthesis increases pH and causes supersaturation of water with bicarbonates (Yates and Robbins 1998). In the pelagic zone the cells of planktonic algae and cyanobacteria (Cyanoprokaryota), particularly their finest fraction, autotrophic picoplankton (APP), were shown to serve as crystallisation nuclei, indispensable in the precipitation of CaCO3 (Dittrich and Obst 2004). In Lake Kierskie, despite the intensive primary productivity and active biological assimilation of CO2, visible in the increased pH value, the amount of CaCO3 collected in the traps between March and April increased only slightly (Fig. 6b, f). The agent blocking precipitation of calcium carbonates were PO43− ions (Kelts and Hsü 1978) abundantly present in the epilimnion (Fig. 6k). Enhanced CaCO3 precipitation started only after the phosphorus ions were assimilated by the primary producers and their concentration in epilimnion decreased. These observations agree with data provided from other lakes (Teranes et al. 1999; Bluszcz et al. 2009). In Lake Kierskie the maxima of carbonates accumulation were observed between mid-April and mid-June (Fig. 6b) and were synchronous with the domination of S. hantzschii, diatom evidenced to have the inductive role in the CaCO3 precipitation in Lake Constance (Stabel 1986). However, Stabel (1986) did not exclude the potential role of many other phytoplankton species in this phenomenon, especially small flagellate forms acting as nucleation sites. Such a species is P. nannoplanctica characterized by the greatest share in the total number of phytoplankton individuals in this time period, particularly in April when the species was not only numerous but also significantly contributed to the biomass production by phytoplankton (Fig. 5; ESM 2).
Oversaturation of waters with CaCO3 was also due to a rapid increase in temperature of the epilimnion (Fig. 6e, h) and degassing of aquatic CO2. However, the maxima of CaCO3 accumulation between 15th April and 25th May (Fig. 6b) result probably not only from enhanced calcite precipitation but also conditions favourable for calcite preservation in the sediments. As observed on 15th April SI values were positive at 16 and 30 m, in contrast to those observed later during the productive season (Fig. 6h). Such conditions prevented carbonates dissolution. Sediment focusing between 15th April and 25th May is evidenced by the higher CaCO3 flux to the sediment traps installed at 30 m compared to the traps at 16 m (Fig. 6b).
In subsequent months, despite the positive calcite SI values in the epilimnion, the amount of CaCO3 flux to the sediment was constantly decreasing (Fig. 6b, h). A drop in CaCO3 precipitation was observed already in May and June and was simultaneous with a decrease in the biomass of S. hantzschii. Phytoplankton dominance shifted towards the species which in the Stabel (1986) opinion do not interfere with the Ca equilibrium in the water. Among them, a large species, Ceratium hirundinella, became dominant. This also concerns the cryptophyte P. nannoplanctica, whose share in the total biomass increased along with the S. hantzschii decrease (Fig. 5; ESM 2). However, based on the gradually decreasing Ca2+ concentration in the epilimnion, negative and constantly decreasing calcite SI values and an increasing Ca2+ concentration in the hypolimnion, especially at 30 m (Fig. 6g, h), it is suggested that the actual CaCO3 flux was greater than measured in the sediment traps as the carbonates were partially dissolved. This suggestion is supported by lower CaCO3 flux to the sediment traps installed at 30 m, i.e. the longer calcite crystals settling time in the water column along with lower water temperatures could have resulted in more CaCO3 being dissolved (Ramisch et al. 1999). Despite the decreased amount of CaCO3 collected, its contribution to the sediment was still the greatest among the basic geochemical components of the sediments (Fig. 6b). Change in the proportion between SiO2 and OM from early spring to the succeeding productive months reflects a shift from the diatom-dominated phytoplankton assemblage to the communities composed of chlorophytes, cryptophytes, and dinoflagellates, the groups with cellulose external covering (Figs. 4c, d, 5).
During the summer stratification rapidly progressing extent of the hypoxic conditions was observed in the hypolimnion (Fig. 4). In September anoxic waters, with O2 concentration below 1 mg l−1, extended between 4 meters below the lake surface and the lake bottom. Such a strong oxygen depletion in the hypolimnion of lake Kierskie is driven by oxic degradation of organic matter supplied from the epilimnion of this highly productive lake. Oxic conditions in the hypolimnion were not improved by aerators installed in the lake that release the oxygenated waters at about 14 m of the water depth and theoretically should increase the oxygen content in the waters. On the other hand, O2 released by aerators may have been instantly used in the process of organic matter decomposition. Although the operation of the aerators in Lake Kierskie is beyond the scope of the present study, the prevailing hypoxic or even anoxic conditions observed in the water column may question their effectiveness. One year of monitoring study does not entitle us to a broader discussion of this issue, however preservation of the laminated sediments in the upper 0.5 m of the sediment sequence (Fig. 2b), though indirect, seems to be a solid record of sustained hypoxia in Lake Kierskie during the winter and summer stratification, despite the 31-years of efforts to oxygenate the waters. The very limited influence of aerators in Lake Kierskie is not an exception. It was shown (Jenny et al. 2016a) that implementation of restoration programs failed to turn off of hypoxia in many European and North American Lakes. Release of the oxygenated waters in the lake hypolimnion can influence geochemical processes in the hypolimnion, including oxic degradation of the organic matter, and therefore modify the sedimentary record. The exact influence of the aerators on the lake and the sediments may be determined only by a detailed analysis of sediment laminations, distinguishing the laminae deposited before and after installation of the aerators. Assuming seasonality of the laminae formation in Lake Kierskie, it is possible to determine the exact lamina deposited in the year when aerators were installed in the lake.
Late summer–early autumn
The early autumn phytoplankton bloom, recorded in the maximum chlorophyll-a values and water transparency restricted to merely 0.9 m (Fig. 6i, j), was observed in mid-September. Water turnover did not start yet, which is indicated by stratification of all the physico-chemical parameters (Fig. 4, ESM 1), therefore nutrients may not have been returned to the epilimnion from hypolimnion. According to data on nutrient loadings in the inflows supplying Lake Kierskie (Grzonka et al. 2016), an increase in TP concentration (0.058 mg l−1 in July, 0.105 mg l−1 in August and 0.152 mg l−1 in September) was observed in the River Krzyżanka, supplying the southern part of Lake Kierskie, where our monitoring study was carried out. This additional nutrient supply could have stimulated the growth of the early autumn phytoplankton assemblage. External sources of nutrients related to human activity, e.g. catchment inputs delivered by inflows, can be highly variable in time and significantly influence the natural processes. The phytoplankton bloom was evidenced in the epilimnion but not in the deeper water layers (Fig. 5; ESM 2) and was caused by the dinoflagellate C. hirundinella that due to large cell sizes and ability to move by means of flagella stays longer in the eplimnetic waters and avoids fast sinking. The biomass produced by phytoplankton in late summer and early autumn indicates highly eutrophic waters, or even hypertrophic if the September biomass value and chlorophyll-a concentrations are considered (Table 1; Fig. 6i). The hypertrophic conditions were not followed by the cyanobacterial dominance, which is rather expected in conditions of high water fertility (Fig. 5; ESM 2). Despite the still high calcite SI values in the epilimnion (Fig. 6h), the early autumn phytoplankton bloom did not result in massive calcite precipitation (Fig. 6b), which was probably prevented by the dominance of C. hirundinella, species regarded as negligible in the CaCO3 precipitation (Stabel 1986). Loss of the sediment traps (stolen between 13th September and 18th October 2016) prevents the determination of the quality and quantity of the sediment accumulated as a result of the late-summer–early-autumn bloom in productivity.
Yearly record of laminated sediments in Lake Kierskie based on the sediment trap study
The most recent laminations in Lake Kierskie are distinct and thick (Fig. 2b). Based on the macroscopic investigation of the short core, the pale, whitish lamina is followed by dark, greyish or black lamina (Fig. 2b). Such a sequence of laminae is confirmed by the sediment flux to the lake bottom recorded by us in a 1-year, monthly sediment trap study (Fig. 6b–d).
Whitish lamina is expected to record chiefly the strong CaCO3 flux from the epilimnion during spring and summer seasons (Fig. 6b). The lamina should start with a layer of centric diatoms that bloomed during the early spring water mixing that resulted in a strong pulse of SiO2 to the sediments (Fig. 6d). The sedimentary record of the early spring phytoplankton bloom is expected to overlap and to be followed by the layer dominated by small calcite crystals, as evidenced by the smear slide analysis of the trap sediments, a record of rapid precipitation from waters oversaturated with respect to calcite. Considering the results of the smear slide analysis, towards the upper limit of the whitish lamina bigger calcite crystals should occur indicating CaCO3 precipitation during late spring and summer when the oversaturation decreased. However, as was shown by Ramisch et al. (1999) in two deep meromictic lakes (288 and 87 m deep), the eventual record of calcite crystals preserved in the sediment may contain a proportionally greater share of larger crystals compared to the size distribution of calcite crystals initially precipitated in the epilimnion. Ramisch et al. (1999) showed that smaller crystals, < 20 um, are easily dissolved and rarely reach the bottom of deep lakes. Although Lake Kierskie is much shallower (35 m in the deepest place), calcite dissolution was proven by increased Ca2+ concentration in the hypolimnion during summer stratification (Fig. 6g) and therefore syn- and postdepositional change in the proportion between small and large calcite crystals must be considered in the detailed examination of the laminae. In lake Kierskie all calcite crystals observed had a typical for calcite crystallographic form of rhombohedra.
The dark (greyish or black) lamina (Fig. 2b) is considered to record sediment deposition during autumn and winter. Following the results of the sediment trap study, it may be subdivided into thicker lamina deposited during or shortly after water turnover, and lamina formed during winter stratification (Fig. 6a). However, this subdivision is based solely on the data from the sediment trap study and must be verified by a microscopic examination of the sediments. The thickness of the lamina deposited during autumn may be an indicator of the water overturn duration in Lake Kierskie, with the thicker lamina being a potential record of prolonged water mixing during homothermy. The sediment accumulated during and after resuspension should be composed of mixed sediment components typical of sedimentation during different seasons, e.g. centric and pennate diatoms and calcite crystals of different sizes. Theoretically, the winter lamina should be the thinnest one and composed primarily of the smallest particles deposited from the suspension. Low sediment flux during winter (Fig. 6a) resulted from the break in the primary productivity in the epilimnion. Stratification of lake waters and discontinuously occurring ice cover protected the deep waters from mixing and sediment resuspension. As showed by the sediment trap study, winter lamina is expected to contain a smaller percentage of calcite crystals in the sediments, a consequence of lack of CaCO3 precipitation in the epilimnion and the dissolution of the carbonates indicated by the negative calcite SI values (Fig. 6h; ESM 1).
Yet another sediment lamina composed of the resuspended material can potentially form during the spring water mixing. However, the material redeposited during the spring lake turnover can be mixed with the flux of the sediment particles from epilimnion, resulting from the bloom of centric diatoms peaking at the water mixis. In spring 2016 an increase in the air temperature and therefore also of Lake Kierskie surface waters was fast (Fig. 6e). Thermal stratification prevented significant resuspension and redeposition of the sediments in the lake.
The results of the one-year monthly monitoring study in Lake Kierskie does not allow us to detect the interannual changes in limnological conditions and sediment flux to the lake bottom as was shown by Bonk et al. (2015), Rull et al. (2017) and Trapote Mari et al. (2018). The planned by us, multi-year monthly monitoring study was interrupted by removal/theft of the traps between 13th September and 18th October 2016. Sediment trap loss was frequently encountered during the CLIMPOL Project (Tylmann, personal communication), therefore our case is not an exception. Considering the location of Lake Kierskie in a densely populated area, its recreational character in the summer, and a year-round presence of fishers on the lake, we did not decide to install the new traps because of a high probability of their repeated loss.
Potential of laminated sediments from Lake Kierskie as a seasonal palaeoenvironmental archive
Lake Kierskie is characterized by a combination of several prerequisites necessary for the formation and preservation of annually laminated sediments (Zolitschka et al. 2015; Tylmann et al. 2013). The study lake is funnel-like shaped with currently significant depth of 35 m (Fig. 1). Considering the 14-m-thick sediment sequence recovered from the deepest part of the lake, the maximum depth of the lake could have been at least 48 m shortly after the lake was formed. Due to prevailing steep slopes in the lakes direct surroundings, the surface area of the lake was not much greater in the past, even if the water level was higher. In consequence, the presently low exposure index (28.3) must have been even lower. In such conditions, the deep waters were even better protected from wind-induced mixing. Moreover, the longitudinal extension of the southern basin of Lake Kierskie is perpendicular to the dominating winds from W and WN direction (Woś 1994). Therefore, during the dominating winds the fetch, and therefore also the wave, is short and the depth of water mixing is restricted.
Despite the potentially favourable morphometric characteristics of the lake and its surroundings, the discontinuous character of laminated sediments was observed in the long core drilled in Lake Kierskie (unpublished data). Such a sediment record indicates that either yearly variability in the seasonal flux to the lake bottom was smaller when the massive, homogenous sediments were deposited, or conditions at the lake bottom did not support the preservation of the laminated sediments. Because the lake is located in the temperate climate with distinct differences between the seasons, influencing primary productivity in the lake, and therefore also sediment flux to the lake bottom, the first hypothesis seems unlikely. Fragments of the sediment sequence where massive deposits are present refer to the times when the record of seasonal signal in the sediments must have been damaged by postsedimentary processes, most likely related to the activity of benthic organisms, presence of which indicates oxic conditions at the lake bottom. The most suitable conditions for varve formation are found in meromictic lakes where the permanent density gradient between the surface and deeper waters develop and the deepest basin of the lake is isolated from mixing and oxygenation (Tylmann et al. 2013), and thus prevents the abundant presence of benthic fauna. Sediments deposited in such conditions have a great potential to record undisturbed seasonal environmental changes. Despite the documented dimictic character of Lake Kierskie and distribution of the oxygenated waters to the lake bottom during the water overturn in the spring and autumn, laminated sediments are well preserved in the top 50 cm of the sediment sequence (Fig. 2b). What prevents the abundant occurrence of the benthic fauna and, thus, protects the most recent laminated sediments from destruction by bioturbations are hypoxic toward anoxic bottom waters during stratification of the lake in the winter and summer (Fig. 4). Similarly, in Lake Montcortès Trapote Mari et al. (2018) observed continuous record of varved sediments despite the interannual shifts between meromictic and mictic states, and temporarily oxic conditions at the lake bottom. In Lake Kierskie the recent strong hypoxia of the hypolimnion is related to the high amount of organic matter supplied from the epilimnion as a result of high nutrient loading to Lake Kierskie controlling the enhanced productivity in the lake. Majority of the lakes with laminated sediments distinguished in NE Poland by Tylmann et al. (2013) were eutrophic, only some mesotrophic. Water hypoxia in response to nutrient loading was shown to be a key factor in the formation and preservation of laminated sediments in lakes across Europe (Jenny et al. 2016a). Concluding, shifts between laminated and massive sediments observed in the long core from Lake Kierskie can indicate changes in the oxygen availability at the lake bottom controlled by the lake trophic conditions. Nevertheless, the formation of the laminated sediments could have been also controlled by other factors affecting the stability of the lake stratification like climatic changes or water level fluctuations (Pleskot et al. 2018; Woolway and Merchant 2019). Therefore, the theoretically undesired in palaeolimnological studies discontinuous character of laminated sediments in Lake Kierskie can be a valuable record of climatic or lake-specific changes, shifting the lake between the laminae preserving and non-preserving states.
Considering the dimictic character of Lake Kierskie the record of seasonal laminae can be destroyed by sediment resuspension induced by the water overturn. However, it has been shown by us that, at least at present, resuspension of the sediment during the water overturn, results in an additional flux of the material, but is not destructive to laminations in the deepest part of Lake Kierskie where the long and short cores were taken.
Potentially, the sediment record of laminated sediments in Lake Kierskie can also be affected by gravity-induced slumping of sediment and turbidity currents. Development of underwater mass movements is favoured by the steep slopes (up to ~ 14°) of the Lake Kierskie southern basin and absence of the flattened surface at the lake bottom. Water saturated sediments are prone to gravitational movements induced by the sediment loading or triggered by waves disturbing the sediments in the shallower parts of the lake (Zolitschka et al. 2015).