Introduction

Chitinous remains of aquatic invertebrates are readily preserved in lake sediments and their stable carbon isotope compositions, expressed as δ13C values, have been used as a proxy to study past changes in the availability and importance of different carbon pathways in lakes (Frossard et al. 2014; Rinta et al. 2016; van Hardenbroek et al. 2010; Wooller et al. 2012). Two such pathways are the fixation of dissolved CO2 by algae and the recycling of carbon from methane (CH4) by methane-oxidising bacteria (MOB). Up to now, palaeolimnological studies that analysed the δ13C values of chitinous invertebrate remains have focused mainly on benthic invertebrates, most prominently chironomid larvae (Insecta: Diptera: Chironomidae), whose chitinous head capsules are regularly found in sediments (Belle et al. 2014; Heiri et al. 2012; van Hardenbroek et al. 2010). Additionally, exoskeleton fragments of planktonic invertebrates such as Bosmina (Branchiopoda: Cladocera: Bosminidae) have been analysed for their δ13C values (Perga 2009, 2011). It has recently been suggested that similar information may be gained from δ13C analysis of resting stages of planktivorous invertebrates (Schilder et al. 2015a, b; van Hardenbroek et al. 2013, 2014; Wooller et al. 2012). Examples of these resting stages are ephippia, produced by Daphnia (Branchiopoda: Cladocera: Daphniidae) and other planktonic cladocerans, and statoblasts, produced by bryozoans that form sessile colonies on hard substrates in the shallow parts of lakes (Wood and Okamura 2005). Ephippia and statoblasts are abundant in lake sediments (Francis 2001; Korhola and Rautio 2001), providing a potential archive for palaeoecological reconstructions.

In modern lake food web studies, the analysis of the δ13C value of different food components is a widely used technique to differentiate between carbon sources (Bunn and Boon 1993). The main organic carbon sources for filter-feeding zooplankton are algae, detritus, and heterotrophic bacteria (Edmondson 1957; Lampert 2011). Using characteristic isotopic signatures of these different sources, carbon flow can be traced through a lake’s food web (Fry 2006; Taipale et al. 2007). Fundamental for such studies is that the carbon isotopic composition of an organism closely reflects the isotopic signature of its diet (DeNiro and Epstein 1978; Peterson and Fry 1987). For Daphnia and their ephippia, Perga (2011) and Schilder et al. (2015b) showed that ephippia δ13C values closely reflect the δ13C values of the living Daphnia at the time of ephippia production. Van Hardenbroek et al. (2016) recently demonstrated that the δ13C values of bryozoan statoblasts are related to the δ13C values of the colonies that produced them. More work is required, however, to show that δ13C values of Daphnia ephippia and bryozoan statoblasts reflect the diet of the parent organisms during or shortly before resting stage formation.

Understanding the extent to which lakes change their capacity to sequester or release carbon in response to global warming and eutrophication is a key issue for palaeoecologists today (Seddon et al. 2014). New insights into lake carbon cycling can be gained from analysing δ13C values of aquatic invertebrate remains. Recently, it was suggested that carbon from CH4 may be an important additional transfer pathway of carbon in lake food webs, indicated by remarkably low δ13C values of aquatic invertebrates when they incorporate CH4-derived, 13C-depleted carbon (Bastviken et al. 2003; Bunn and Boon 1993; Grey 2016). The δ13C values of biogenic CH4 in small European lakes range between −86 and −61‰ (Rinta et al. 2015), which is markedly depleted in 13C relative to algae with δ13C values that typically fall in the range from −35 to −25‰ (Jones et al. 1999; Peterson and Fry 1987). The large difference in δ13C values thus allows for differentiation between photosynthetically produced organic matter and MOB as potential carbon sources for primary consumers in lakes. Significant relationships have been observed between δ13C of Daphnia ephippia in surface sediments and diffusive CH4 flux (Van Hardenbroek et al. 2013) and within-lake CH4 concentrations (Schilder et al. 2015a). Cladoceran remains may thus record changes in past diffusive CH4 flux in lakes, though they do not provide direct information on other forms of CH4 fluxes such as ebullition or plant-mediated transport.

Palaeoecological studies using δ13C values of chitinous remains have largely overlooked the use of bryozoan statoblasts, which have been suggested as a resource for reconstructing δ13C values of primary production (Turney 1999; van Hardenbroek et al. 2014, 2016). As filter feeders, Bryozoa rely on algae, particulate organic matter (POM) and associated microorganisms (e.g. bacteria) as their main food source (Kaminski 1984). Bryozoa are mainly bound to their substrate and do not have access to food sources from deeper water layers and are therefore not, or only to a small degree, influenced by CH4-derived carbon (van Hardenbroek et al. 2016). By combining δ13C measurements of mobile filter feeders like Daphnia with sessile filter feeders like Plumatella (Phylactolaemata: Plumatellida: Plumatellidae), it is possible to separate changes in the importance of CH4-derived carbon from changes in algal δ13C values, or more generally, changes in lake productivity (van Hardenbroek et al. 2014; Rinta et al. 2016). At present, however, no information is available about seasonal changes in δ13C values of cladoceran and bryozoan resting stages or about the extent to which this seasonality influences the δ13C values of statoblast and ephippia remains in lake sediments. Furthermore, it is unknown whether the δ13C values of deposited organism remains reflect integrated δ13C values for statoblasts and ephippia that float on the lake surface, or whether the sedimentary assemblages are characterized by systematically higher or lower values. Such differences could appear as a consequence of degradation, transport processes, or production of resting stages in cryptic habitats that do not contribute to floating resting stages at the lake surface (e.g. from deeper littoral, lower epilimnetic or profundal habitats). These uncertainties currently hinder the development of δ13C analysis of invertebrate resting stages as a proxy for reconstructing past variations in carbon cycling in lakes.

The main aim of this study was to assess the influence of seasonality on δ13C values of zooplankton remains and evaluate their potential use as palaeoecological indicators for lacustrine carbon cycle dynamics. This study provides a first assessment of seasonal variations in δ13C values of cladoceran and bryozoan resting stages in a small temperate European lake, with respect to physical, chemical and biological variables. In particular, the focus was on assessing seasonal variations in transfer pathways of carbon in the lake, and their influence on the δ13C values of organisms in the water column, as well as their chitinous resting stages (i.e. planktonic Cladocera and their ephippia, bryozoan statoblasts). In a field campaign, δ13C values of Daphnia and their ephippia were analysed over a 2-year period, and δ13C values of floating resting stages of Ceriodaphnia (Branchiopoda: Cladocera: Daphniidae) and Plumatella were analysed when present during this period. In a second step, we investigated the implications of seasonal changes for interpretation of δ13C values in fossil invertebrate remains. To that end, ephippia of the cladocerans Daphnia and Ceriodaphnia, and statoblasts of the bryozoan Plumatella were analysed in the lake surface sediment and in a short sediment core covering roughly the past 150 years. Our study was conducted on Lake Gerzensee (7°33′E, 46°50′N, 606 m a.s.l.), a small temperate lake in the foreland of the Swiss Alps, about 20 km south of Bern (Fig. 1). The lake is characterised by exceptionally high lake water CH4 concentrations (Rinta et al. 2015; Schilder et al. 2016). Lake Gerzensee has a surface area of 0.27 km2, a total water volume of 0.16 km3, and a maximum depth of 10 m. The catchment area (2.6 km2) consists of 80% agricultural land, 5% wooded land, and 15% urban areas (Lotter et al. 2000). The mean annual temperature is 8.8 °C (Bern Zollikofen 1981–2010, Bundesamt für Meteorologie und Klimatologie MeteoSchweiz 2014). Today, Lake Gerzensee is eutrophic, with summer anoxia in the hypolimnion (Zeh et al. 2004).

Fig. 1
figure 1

Map of Switzerland showing the location of Lake Gerzensee in the Swiss foreland of the Alps (triangle) and the bathymetry of the lake. Black circles indicate the fieldwork measuring station (C1) and coring site

Full size image

Materials and methods

Lake Gerzensee was sampled on 15 one-to-three-day visits between October 2012 and July 2014, throughout all seasons (Electronic Supplementary Material [ESM] Table S1). In the first period, between October 2012 and September 2013, measurements were taken every other month (6 visits). Based on the first year of data, the fieldwork protocol was adjusted to gain more detailed information about deeper-water POM and chlorophyll a concentrations (ESM Table S1). In the second year, intervals between field visits were reduced to two weeks from September 2013 to December 2013, to cover the destratification period (4 visits), and sampling continued with 2–4 week intervals after the lake ice thawed in February 2014, until July 2014 (5 visits). Data for March and July 2014 were collected along with other fieldwork activities, using a shortened protocol (ESM Table S1).

Sample collection

Vertical profiles of temperature and dissolved oxygen concentration ([O2]aq) were measured for every metre in the water column at the lake centre (location C1, Fig. 1), using a multi-sensor probe (WTW CellOx© 325 oxi1970i, Germany). Daphnia individuals were collected from the oxic part of the water column at the lake centre (location C1) in multiple vertical hauls, with a 40-μm mesh plankton net. Flotsam was collected from the lake surface with a hand net (mesh size ~120 μm). Water samples were collected at C1 in 0.7 and 8 m water depth (top epilimnion and bottom hypolimnion, respectively), using a 5-L water sampler (UWITEC, Austria). For these water samples, pH was measured (Waterproof pHTestr 20, Oakton, USA), and 60 ml of water was injected with a syringe into a 118-ml glass vial through a 10-mm-thick butyl rubber stopper (Apodan, Denmark) to determine the abundance and δ13C value of the dissolved inorganic carbon (DIC). The vials were prepared beforehand with 200 μl of H3PO4 (85%), closed, and repeatedly vacuumed and flushed with N2 to ensure that no CO2 remained in the vials (Rinta et al. 2015; Schilder et al. 2015a). Samples for the δ13C analysis of POM were collected from the lake water at C1 (0.7 m water depth) by passing the water through a 250-μm sieve before manually pushing water through a glass fibre filter (Whatman GF/C 25 mm, pore size 1.2 μm) with a syringe (water volume recorded in the field). A second sample of POM was obtained in a similar fashion, immediately put into 90% undenatured ethanol, and covered with aluminium foil for chlorophyll a analysis. Immediately after returning from fieldwork, POM filters designated for chlorophyll a analysis were heated to 70 °C in 90% undenatured ethanol for 10 min, put in an ultrasound bath for 5 min, and stored in a refrigerator for at least 48 h. The samples were then filtered through a membrane filter to remove suspended particles. Chlorophyll a measurements were based on absorbance at characteristic wavelengths (Schwoerbel 1994). Samples were injected into glass cuvettes (1 cm light path) and absorbance at wavelengths (λ) 665 nm and 750 nm was measured with a spectrophotometer (Jenway, UK). Chlorophyll a content in the lake water was calculated following EAWAG (1995).

Isotope analysis

δ13C analysis of DIC and POM

The concentration of CH4 and CO2 in the headspace of the DIC samples was measured by gas chromatography with a flame ionisation detector and methanizer (GC-FID; Shimadzu GC8, PoropackN column, see Rinta et al. 2015 for details). Lake water DIC concentrations and dissolved CH4 concentrations ([CH4]aq) were back-calculated from measured headspace CO2 and CH4 concentrations. Dissolved CO2 concentrations ([CO2]aq) (as the sum of dissolved CO2 and H2CO3) were calculated following Stumm and Morgan (1996), accounting for lake water pH, temperature, and DIC concentrations.

Filters containing lake water POM were freeze-dried and a maximum amount of filter material was separated from the seston. The seston was then transferred into ultra-clean tin cups. For δ13C analysis of DIC, 15–20 ml of gas from the headspace of the vial was allowed to escape into a 60-ml syringe. The gas was then injected into a pre-vacuumed 12-ml glass vial such that slight overpressure was applied. The procedure was repeated for δ13CH4 measurements. Isotope samples of CO2, CH4, and POM were analysed at the Stable Carbon Isotope Facility of the University of California-Davis on an Elementar Vario EL Cube or Micro Cube elemental analyser interfaced to a PDZ Europa 2020 isotope ratio mass spectrometer (IRMS). Analytical uncertainties for gas sample δ13C values were <0.1‰ (one standard deviation) for CO2 for two to three replicate measurements of three laboratory standards (δ13C = −40.73, −10.39, and −3.59‰) and <0.1‰ (one standard deviation) for CH4 for replicate measurements of a laboratory standard (δ13C = −36.7‰, n = 6). Results are reported in conventional δ-notation relative to the international standard Vienna PeeDee Belemnite (V-PDB).

δ13C analysis of zooplankton

Living Daphnia were kept in approximately 1 L of unfiltered lake water for 1–2 days. They were then separated from other organisms under a dissecting microscope (magnification 20×–50×), and were frozen in demineralised water. Samples were freeze-dried and weighed into ultra clean tin cups (Lüdi Swiss AG, Switzerland). Because Daphnia ephippia cannot be identified to species level under the microscope, the species of living Daphnia was not determined for this analysis. Flotsam was sieved at 100 μm and examined under a dissecting microscope. Ephippia of the planktonic cladocerans Daphnia and Ceriodaphnia, and statoblasts of the bryozoan Plumatella were the only invertebrate resting stages found regularly and abundantly enough for δ13C analysis. These resting stages were identified according to Vandekerkhove et al. (2004) (Cladocera) and Wood and Okamura (2005) (Bryozoa), exposed to 10% potassium hydroxide (KOH) for 2 h (van Hardenbroek et al. 2010), rinsed 5–10 times with demineralised water, and picked into ultra-clean tin cups. Isotope samples of Daphnia as well as ephippia and statoblasts were also analysed at the Stable Carbon Isotope Facility of the University of California-Davis on an Elementar Vario EL Cube or Micro Cube elemental analyser interfaced to a PDZ Europa 2020 IRMS. Sample sizes were in the range of 120–250 individuals for Daphnia (150–500 μg), 150–200 for Daphnia ephippia (150–250 μg), 200–300 for Plumatella statoblasts (150–300 µg), and 450–600 (150–200 μg) for Ceriodaphnia ephippia. Analytical uncertainties for invertebrate δ13C measurements were ≤0.6‰ (one standard deviation) for replicate measurements (n = 3–36) of five laboratory standards (Bovine Liver (δ13C = −21.7‰), USGS-41 Glutamic Acid (δ13C = 37.6‰), Nylon 5 (δ13C = −27.7‰), Peach Leaves (δ13C = −26.1‰), Glutamic Acid (δ13C = −28.9‰)).

Sediment analysis

Sediment coring and chronology

In October 2012, a 62-cm-long sediment core (GER12) was recovered from the centre of Lake Gerzensee (9.5 m water depth) using a gravity corer (UWITEC, Austria). Upon arrival in the laboratory, core GER12 was sampled at 1-cm intervals and freeze-dried until further analysis. For 210Pb and 137Cs dating, a total of 15 freeze-dried samples from the upper 48 cm of the core were analysed using gamma spectrometry at the Department of Chemistry and Biochemistry at the University of Bern, Switzerland. 210Pb (46.5 keV), 241Am (59.5 keV), 226Ra progenies (351.9 and 609.3 keV), and 137Cs (661.7 keV) were measured using a Broad Energy Germanium (BEGe) Canberra detector with low background and high absolute full-energy peak efficiencies for close on-top geometries of >20% and ~5% for 210Pb and 137Cs, respectively.

δ13C analysis of sedimentary invertebrate remains

For invertebrate δ13C analysis, core GER12 was sampled every fourth centimetre. When sample mass of the invertebrate remains was not sufficient for δ13C analysis, the lower adjacent centimetre was added to the sample. Nonetheless, some samples in the lower half of the core had to be pooled with the next regular-interval sample to obtain a sufficient number of remains for analysis. Hence, invertebrate δ13C values represent remains from up to 6 cm of sediment. For each sample, 50% by weight of the freeze-dried material was deflocculated in 10% KOH for 2 h and sieved at 100 μm (van Hardenbroek et al. 2010). Daphnia ephippia, Ceriodaphnia ephippia, and Plumatella statoblasts were identified according to Vandekerkhove et al. (2004) for Cladocera and Wood and Okamura (2005) for Bryozoa and separated from the sediment. Remains were treated with 2 M NH4Cl solution buffered with 0.35 NaOH for 20 h to remove carbonates (Verbruggen et al. 2010), and picked into pre-weighed silver cups (6 × 4 mm; Säntis, Switzerland). All fossil invertebrates were analysed at the Alaska Stable Isotope Facility of the University of Alaska, Fairbanks, on a Costech ESC 4010 elemental analyzer interfaced via a ThermoConflo III to a Thermo Delta V IRMS. Analytical uncertainties were <0.1‰ for replicate measurements of a laboratory standard (peptone, δ13C = −15.8‰, n = 24). Sample sizes were in the range of 20–55 remains for Daphnia ephippia and Plumatella statoblasts, and 100–200 for Ceriodaphnia ephippia. In addition to the sediment core, material from a sediment trap located close to the coring site at the centre of Gerzensee during 2012/13 was available from another study (C. Adolf, unpublished). The trap was placed 3 m above the lake floor and remained in the lake for 12 months. The material was processed in the same way as the sediment core, and one sample for Ceriodaphnia ephippia was analysed for its δ13C value.

Statistical analysis

Statistical analyses on material collected alive and as floatsam were performed in R (R Development Core Team 2008). For Pearson correlation tests, normality of the variables was tested prior to analysis using the Shapiro–Wilk test. When a normal distribution of the data was rejected, Spearman’s rank correlation was used to test for correlations. The average value reported for fieldwork parameters was calculated as a weighted mean of all samples collected during the campaign and each sample was weighted by the time interval for which it is most representative.

Results

Physical and chemical lake conditions

All variables showed characteristic seasonal variations throughout the two years of monitoring. Measured surface water temperatures ranged between 4.6 °C in December 2013 and 25.8 °C in July 2013, with a mean of 12.4 °C (Fig. 2a). The bottom water (8 m depth) was anoxic ([O2]aq < 1 mg L−1) from June to November (Fig. 2b). Chlorophyll a values ranged from 15.2 μg chl a L−1 in autumn 2013 to 7.2 μg chl a L−1 in February 2014, but only one measurement was taken during summer (July 2014). Deep-water chlorophyll a measured at 7 m depth in May and July 2014 was comparable (July, 14.7 μg chl a L−1) or higher (May, 14.2 μg chl a L−1) than surface water measurements (12.6 and 8.6 μg chl a L−1, respectively). [CO2]aq in the epilimnion ranged from peak values just after autumn mixing and in spring (96.7–192.8 μM) to 9.6 μM in September (ESM Fig. S1). [CH4]aq in the water column reached high values (1082 μM) in the hypolimnion during summer stratification. Surface water [CH4]aq was highest in November during autumn turnover (49.8 μM, ESM Fig. S1).

Fig. 2
figure 2

a Temperature and b dissolved oxygen concentrations ([O2]aq) measured at 1-m depth intervals in the water column at the lake centre (location C1) over the fieldwork period in 2012–2014. (Colour figure online)

Full size image

Stable carbon isotopes

With the exception of Daphnia ephippia, all measured variables showed seasonal variation in δ13C values, with the lowest values right after autumn lake mixing and throughout winter, and the highest values in summer (Fig. 3a).

Fig. 3
figure 3

a δ13C values of CO2, particulate organic matter (POM), Daphnia for location C1 (pelagic zone) and Daphnia ephippia (whole lake) over the fieldwork period. b δ13C values of Daphnia ephippia, Ceriodaphnia ephippia, and Plumatella statoblasts (whole lake) over the fieldwork period. Daphnia and flotsam δ13C values represent single measurements except for 25 February 2014 and 27 March 2014 when enough material was available for three measurements of Daphnia δ13C. In those cases average values are presented

Full size image

The isotopic composition of [CO2]aq \( \left( {\updelta^{13} {\text{C}}_{{{\text{CO}}_{2} }} } \right) \) for location C1 at the lake surface was most 13C-depleted in February and March (\( {\updelta^{13} {\text{C}}_{{{\text{CO}}_{2} }} } \) ~ −21‰), whereas the highest \( {\updelta^{13} {\text{C}}_{{{\text{CO}}_{2} }} } \) value of −12.1‰ was recorded in July 2013 (Fig. 3a). The seasonal cycle in \( {\updelta^{13} {\text{C}}_{{{\text{CO}}_{2} }} } \) values was less pronounced in the hypolimnion (data not shown). During summer, a vertical gradient in \( {\updelta^{13} {\text{C}}_{{{\text{CO}}_{2} }} } \) values developed between the epilimnion and the hypolimnion, which was eliminated at lake mixing in autumn. At the lake centre, the δ13C value of surface water POM (δ13CPOM) was highest during summer (−28.5‰), and reached its minimum in March 2013 (−41.6‰, mean = −34.7‰; Fig. 3a). Bottom water δ13CPOM, measured at 7 m depth, was in the range of surface water δ13CPOM for measurements in April and May 2014 (ESM Fig. S2d). However, during summer stratification in July 2014, bottom water δ13CPOM values were distinctly more negative than surface water δ13CPOM (−38.0 and −30.0‰, respectively). Surface water δ13CPOM was on average 17.0‰ (range 13.2–21.0‰) more negative than surface water \( {\updelta^{13} {\text{C}}_{{{\text{CO}}_{2} }} }\), with the largest differences measured in March 2013 and May 2014 (Fig. 3a).

Daphnia δ13C values (δ13CDaph) in the lake centre were on average 3.4‰ more negative than δ13CPOM values (Fig. 3a). The average δ13CDaph value was −39.4‰. Values ranged from −44.2‰ in early spring to −29.8‰ in summer. δ13CDaph values were positively correlated with chlorophyll a concentrations in surface water POM (Pearson correlation r = 0.86, p < 0.05, n = 6; ESM Fig. S2a), and [CH4]aq in bottom waters (log-transformed, Pearson correlation r = 0.86, p < 0.01, n = 13; ESM Fig. S2b). Moreover, bottom water [CH4]aq showed a negative correlation with the offset between δ13CPOM and δ13CDaph (Δ13CPOM–Daph) (Spearman’s rank correlation r = 0.67, p < 0.05, n = 13; ESM Fig. S2c).

In contrast to δ13CDaph values, the δ13C values of floating Daphnia ephippia (δ13CDaphFlot) did not show seasonal variations, but instead remained relatively constant around −39.5‰ (range −41.7 to −38.8‰, Fig. 3a, b). Interestingly, floating Ceriodaphnia ephippia δ13C values (δ13CCerioFlot) showed changes over time, with the lowest δ13C values down to −50.1‰ in winter (Fig. 3b). On average, δ13CCerioFlot was more 13C-depleted (mean = −42.9‰, range −50.1 to −34.3‰) than δ13CDaphFlot, whereas Plumatella statoblast δ13C (δ13CPluFlot) was least 13C-depleted (mean = −36.3‰, range −40.0 to −34.6‰). In-lake abundance of the floating remains could not be assessed by the applied sampling technique, and no structural degradation of the remains was visible under the microscope on any of the fieldwork dates. Ceriodaphnia ephippia from the sediment trap had a δ13C value of −42.4‰.

Sediment core and chronology

For core GER12, activity of unsupported 210Pb was transformed into an age-depth model for the upper 48 cm using the constant rate of supply (CRS) model (Appleby 2001; Appleby and Oldfield 1978) (Fig. 4). Activity measurements of 137Cs showed a distinct peak at 28.5 cm and a second smaller peak at 20.5 cm. Many European lakes show two maxima in 137Cs activity, which are associated with aboveground nuclear bomb tests in AD 1963 and the Chernobyl reactor accident in AD 1986 (Appleby 2001). For Swiss lakes, the latter peak is usually greater than the first (Albrecht et al. 1998; Lotter et al. 1997). This was not the case in Gerzensee sediments where only one measurement point defined the anticipated Chernobyl peak. The Chernobyl peak may thus not be fully revealed in the record because of the relatively low sampling resolution, i.e. every 4 cm. Therefore the 210Pb model was not modified to fit the 137Cs profile. The lower peak in 137Cs activity, expected to coincide with the nuclear bomb peak in AD 1963, was confirmed by 241Am and agrees well with the 210Pb chronology (Fig. 4).

Fig. 4
figure 4

Age-depth model (black line, triangles) for core GER12 based on 210Pb activity. The grey line shows the accumulation of 137Cs activity in the record (based on the 210Pb-inferred accumulation rate), with the diamond showing the location of the lower activity peak of 137Cs in the 210Pb-based age model. This peak is correlated with atmospheric nuclear bomb testing in AD 1963 and coincides with the only measurable 241Am activity in the core (1.9 ± 0.2 Bq/kg)

Full size image

The abundance of invertebrate remains in the sediments of Gerzensee allowed for the analysis of eight samples of Daphnia ephippia covering the upper 27 cm, nine samples of Plumatella statoblasts (0–43 cm), and 15 samples of Ceriodaphnia ephippia (0–62 cm). The δ13C values of all three invertebrate genera showed only small variations throughout the record (Daphnia ephippia −39.8 to −37.9‰, Ceriodaphnia ephippia −43.3 to −39.2‰, and Plumatella statoblasts −33.3 to −29.7‰; Fig. 5). δ13C values of Plumatella statoblasts were on average distinctly less negative (−32.0‰) than the remains of the two mobile filter feeders Daphnia and Ceriodaphnia (−39.0 and −41.9‰, respectively). The most negative δ13C values were measured for Ceriodaphnia ephippia with the minimum of −43.3‰ at 41–42 cm depth (AD ~ 1920). Remains of Daphnia and Ceriodaphnia showed a very similar pattern, with one distinct peak of less negative δ13C values at about 18 cm depth (AD ~ 1980). In contrast, δ13C values of Plumatella statoblasts remained constant during this period, but show a 3‰ increase around 30 cm (AD ~ 1955). Besides differences in δ13C values, Plumatella statoblasts had a lower C:N ratio (mean = 4.3) compared to values for Daphnia and Ceriodaphnia ephippia (6.3 and 6.5, respectively).

Fig. 5
figure 5

δ13C values of fossil invertebrate remains of Daphnia, Ceriodaphnia and Plumatella in sediment core GER12 and from the flotsam of the lake (flotsam values plotted above 0 cm sediment core depth). Symbols in the downcore record represent the average depth of the measurements in cases for which samples were pooled to obtain sufficient weight for δ13C analysis. Measurements may represent remains from up to 6 cm of sediment (see text for details)

Full size image

Discussion

Seasonal variation in Daphnia δ13C values

Daphnia δ13C values were consistently below −38.5‰ and reached minimum values, less than −44‰, in the period from autumn to spring. This is well below commonly reported values for algae, which usually lie above −35‰ (France 1995; Peterson and Fry 1987; Vuorio et al. 2006). Nonetheless, δ13CDaph values closely track the seasonal cycle of surface water δ13CPOM values (Fig. 3a), but with an average difference of 3.4‰ between δ13CPOM and δ13CDaph. Hence, Δ13CPOM–Daph was slightly higher than values reported by del Giorgio and France (1996) for zooplankton in Canadian lakes and other published values discussed by these authors (mean difference 2.6‰). A positive correlation between δ13CDaph values and surface water chlorophyll a concentrations, which can serve as a proxy for algal concentrations (Sartory and Grobbelaar 1984), was observed between autumn and spring (ESM Fig. S2a). Hence, low δ13CDaph values coincide with relatively low concentrations of algae in surface water POM. During times when algae are less abundant, other food sources may contribute to the carbon uptake of Daphnia (Taipale et al. 2008). Several studies have shown that Daphnia and other invertebrates can incorporate CH4-derived carbon to supplement their diet (Bastviken et al. 2003; Devlin et al. 2015; Kankaala et al. 2006). In a study of a small Finnish lake, Taipale et al. (2008) found that MOB contributed to Daphnia’s diet throughout the year, comprising up to 50% of the diet in autumn. Considering the high CH4 concentrations in Gerzensee, uptake of 13C-depleted carbon by feeding on MOB seems the likely explanation for the exceptionally low δ13CDaph values measured in Gerzensee.

The relationship between Daphnia δ13C values and [CH4]aq

Field studies that compared the carbon isotopic composition of invertebrate remains in surface sediment samples with in-lake CH4 abundance suggest that a relationship exists between the δ13C values of some invertebrate groups and lake water CH4 concentrations (Schilder et al. 2015a; van Hardenbroek et al. 2013). Within our 2-year measurement period at Gerzensee, δ13CDaph was positively correlated with [CH4]aq in the bottom waters, i.e. we recorded δ13CDaph values more similar to δ13CPOM during times of high bottom water [CH4]aq (ESM Fig. S2b). This is also indicated by the negative correlation between [CH4]aq in the bottom waters and Δ13CPOM–Daph (ESM Fig. S2c). Hence, during lake stratification, when [CH4]aq is increasing in the hypolimnion, Daphnia utilise carbon whose isotopic signature is more similar to δ13CPOM than at times when the water column is mixed. This suggests that during lake stratification, algae are the main food source of Daphnia, whereas Daphnia may rely more heavily on other food sources when the lake is mixed. In a multi-lake study, Schilder et al. (2015a) found a negative correlation between late summer [CH4]aq in both bottom and surface waters, and δ13C values of Daphnia ephippia isolated from surface sediment samples, indicating that Daphnia are more likely to incorporate CH4-derived carbon in lakes with high CH4 abundance. With regard to this relationship, our results imply that in lakes with high CH4 accumulation during summer stratification, more 13C-depleted carbon becomes accessible to the food web upon mixing and oxygenation of CH4 in the water column, and this signal is seen in the floating ephippia. Other processes, e.g. increased transport of allochthonous organic carbon and higher associated CH4 production may also contribute to the importance of CH4-derived carbon for Daphnia in the autumn months.

Floating cladoceran and bryozoan remains

In a field study, Perga (2011) showed that there is no significant carbon isotope fractionation between Daphnia and their ephippia. This was confirmed in a laboratory experiment by Schilder et al. (2015b), which showed that Daphnia ephippia δ13C values closely resemble those of Daphnia during ephippia production under different environmental conditions (−0.2 ± 0.4‰ for 12 °C, 1.3 ± 0.3‰ for 20 °C). In Gerzensee, δ13CDaphFlot remained very similar across the annual cycle (Fig. 3a, b). If no fractionation is assumed during ephippia production, this suggests that the ephippia were produced batch-wise and then floated on the lake for several months. In the data set presented here, periods when δ13CDaphFlot coincided with δ13CDaph include December 2012, June 2013, November 2013, and April 2014. This is in line with the observation that ephippia production commonly occurs in early summer and late autumn (Cáceres 1998), when environmental conditions deteriorate, e.g. there is temperature decline, crowding, oxygen depletion, or limited food availability (Korhola and Rautio 2001). Hence, in Gerzensee, changes in δ13CDaphFlot reflect δ13CDaph during ephippia production, i.e. in late autumn and/or early spring, but δ13CDaphFlot does not reflect seasonal changes in the diet of Daphnia. This observation differs from the findings of Schilder (2014) at Dutch Lake De Waay, where Daphnia ephippia followed the seasonal variation of Daphnia δ13C, indicating several production periods per year. The reason for the different timing of ephippia production is unclear, although it could be explained by differences in species composition between the two lakes. Regardless of the cause, our results indicate that different production intervals, and possibly switching between different production intervals over time, potentially influences Daphnia ephippia δ13C values in sediment records.

Floating Ceriodaphnia ephippia show larger variability in δ13C values than Daphnia ephippia (Fig. 3b). To our knowledge, no prior study has assessed the fractionation between maternal Ceriodaphnia and the chitinous structure of their resting eggs. Thus, following the simplest assumption of no (or constant) fractionation, as is observed for Daphnia, our results suggest that Ceriodaphnia ephippia are produced at several times or continuously throughout the year. No living Ceriodaphnia were analysed in this study, but the broad range of δ13CCerioFlot values (−34.3 to −50.1‰), which is comparable in span to the seasonal cycle of living Daphnia (−29.8 to −44.2‰), may be an indication that the changes in δ13CCerioFlot represent (a part of) the seasonal cycle of Ceriodaphnia δ13C.

Floating Plumatella statoblast δ13C values also show more pronounced seasonal changes than δ13CDaphFlot, but the variability is not as large as for δ13CCerioFlot (Fig. 3b). Since no living Plumatella zooids were analysed, no pattern of statoblast production can be determined. However, a recent study by van Hardenbroek et al. (2016) suggests that the δ13C values of Plumatella statoblasts collected from colonies are significantly correlated with Plumatella zooid δ13C. Okamura and Hatton-Ellis (1995) found that production of statoblasts may start in mid-summer, but highest production occurs in late summer and early autumn. Overwintering statoblasts are released when colonies collapse in late autumn. Following these findings, δ13C values of floating Plumatella statoblasts may reflect zooid δ13C values during statoblast production from late summer to late autumn.

Linking δ13C values of floating resting stages to those of fossil invertebrate remains

The average δ13C value of floating Daphnia ephippia (−39.5‰ ± 0.9 [one standard deviation]) is in excellent agreement with the δ13C value of Daphnia ephippia in the uppermost 1–3 cm of the sediment at the centre of the lake (−39.8‰; Fig. 5), and with the average δ13C value of living Daphnia collected during the 2-year fieldwork period (−39.4‰ ± 4.6). This confirms that δ13C values of Daphnia ephippia in the sediment of Gerzensee reflect the δ13C signal of floating Daphnia ephippia during the year. Seasonal changes in δ13CDaph, however, are not recorded by floating or deposited Daphnia ephippia δ13C values in the lake (see previous sections). This has implications for the interpretation of the sediment core record of Daphnia ephippia δ13C values. In Gerzensee, nearly constant δ13CDaphFlot values throughout the year indicate batch-wise production of ephippia in autumn and/or early spring. Changes in the δ13C value of sedimentary Daphnia ephippia may thus record changes in the importance of MOB in the diet of Daphnia (only) during spring and/or autumn. The latter period is indirectly coupled with the accumulation of [CH4]aq during summer stratification, which is the reason for high CH4 abundance in the water column during autumn lake mixing. Hence δ13C values of sedimentary Daphnia ephippia may indicate the accumulation of [CH4]aq during summer stratification in small European lakes like Lake Gerzensee, as suggested by Schilder et al. (2015a).

The Ceriodaphnia ephippia δ13C value recorded in the sediment trap during 2012/13 (−42.5‰) is in excellent agreement with the surface sediment measurement presented here (−42.9‰, 1–3 cm; Fig. 5). The average flotsam δ13C value (−42.9‰ ± 4.8) also agrees very well with the surface sediment measurement, although no full annual cycle is covered by the flotsam measurements (October to May only, Fig. 3b). Our study is the first to assess and confirm that the δ13C values of Ceriodaphnia ephippia in the sediment can provide information about the δ13C value of floating ephippia at the time of deposition.

The Plumatella statoblast δ13C value in the surface sediment (−32.7‰, pooled 1–3 and 5–7 cm, this study) was distinctly less negative than the average δ13C value of floating statoblasts (−36.3‰ ± 1.7; Fig. 5). It should, however, be noted here that summer measurements of δ13CPluFlot are underrepresented in our dataset, with only one measurement in July 2013 (Fig. 3b). It therefore remains uncertain whether a better representation of the summer period may resolve this mismatch between average flotsam values and the surface sediment measurement.

Interpreting fossil invertebrate δ13C values

Daphnia ephippia δ13C values in our sediment record (−39.8 to −37.9‰) were continuously below −35‰, whereas algae are commonly reported to have δ13C values above −35‰ (France 1995; Peterson and Fry 1987; Vuorio et al. 2006). Our values are comparable to sedimentary Daphnia ephippia δ13C values found in Lake De Waay, the Netherlands (Schilder 2014), but are distinctly more negative than cladoceran δ13C values reported from sediment records of Lake Strandsjön, Sweden (van Hardenbroek et al. 2014) and Lake Annecy, France (Frossard et al. 2014). In sediment records covering longer time periods, van Hardenbroek et al. (2013) and Wooller et al. (2012) found δ13C values over the range of all of the studies mentioned above. In a recent study in Lake Mekkojärvi, Finland, Rinta et al. (2016) showed that Daphnia ephippia δ13C values changed abruptly from values below −45‰ to values greater than −40‰, shifts that these authors interpreted as representing changes in the availability of CH4 in this lake. As discussed earlier, δ13C values in Daphnia ephippia have been shown to correlate negatively with CH4 abundance in the hypolimnia of small European lakes (Schilder et al. 2015a). Therefore, the observed Daphnia ephippia δ13C values below −35‰ suggest that [CH4]aq remained high at Gerzensee throughout the record. Only the least negative δ13C values, around 20 cm (AD ~ 1970–1980), could indicate a short period with a reduced influence of 13C-depleted carbon in the diet of Daphnia, but δ13C values remained clearly more negative than commonly reported algal δ13C values. Lower Daphnia δ13C values observed for the sediments of eutrophic and stratified lakes Gerzensee and De Waay can potentially be explained by a greater importance of CH4-derived carbon compared to that in less nutrient-rich Lake Annecy and non-stratified Lake Strandsjön, where algae are the main carbon source of Daphnia. Mekkojärvi is a very small, stratified, humic lake with low oxygen concentrations within 1 m of the lake surface (Rinta et al. 2015), which may explain the high relevance of CH4-derived carbon in its planktonic food web.

Ceriodaphnia ephippia were not analysed in any of the above-mentioned studies. Ceriodaphnia ephippia, however, had even more negative δ13C values than Daphnia in the Gerzensee sediment (range −43.3 to −39.4‰; Fig. 5). This indicates a similar, but potentially more dominant source of 13C-depleted carbon for Ceriodaphnia ephippia. The systematic 13C-depletion of Ceriodaphnia ephippia relative to Daphnia ephippia may be caused by a difference in the average particle size that these taxa filter from the water. Ceriodaphnia feed on smaller particles than Daphnia (Geller and Müller 1981), potentially containing a higher proportion of MOB and other bacteria. The relative 13C-depletion of Ceriodaphnia could also be caused by differences in body composition, e.g. lipid content, between taxa, as has been suggested by Matthews and Mazumder (2005) for other zooplankton species. However, C:N ratios, which may be an indicator of lipid content (Matthews and Mazumder 2005), do not show large differences between Daphnia and Ceriodaphnia ephippia (mean = 6.3 and 6.4 for the uppermost 27 cm of the sediment record, respectively). Therefore, low δ13C values throughout the record suggest that, similar to Daphnia, Ceriodaphnia rather consistently incorporated a 13C-depleted carbon source over the past 150 + years, at least during the season(s) of ephippia production.

Plumatella statoblast δ13C values in our sediment record were distinctly less negative compared to Daphnia and Ceriodaphnia ephippia (−33.3 to −29.7‰; Fig. 5). This suggests that Plumatella zooids accessed carbon sources that were less 13C-depleted compared to those on which the two cladoceran taxa relied. Similar δ13C values were reported by van Hardenbroek et al. (2014), for Plumatella statoblasts in shallow Lake Strandsjön in Sweden (−33.1 to −28.0‰). Our δ13C values for Plumatella statoblasts are well in the range of commonly reported δ13C values for algae. Hence, it appears that δ13C values of sedimentary Plumatella statoblasts can potentially provide information about the temporal evolution of algal δ13C, and thus help track past changes in the lake’s carbon cycle. However, a single statoblast flotsam sample from our study, collected during autumn mixing in November 2013 (Fig. 3b), yielded a δ13C value of −40.0‰, well below the values expected for algae in Lake Gerzensee. Similarly, individual bryozoan samples with very negative δ13C values have been reported in previous studies (e.g. Rinta et al. 2016; van Hardenbroek et al. 2016). This suggests that these organisms may also ingest 13C-depleted carbon sources such as MOB under circumstances when they are abundant in their habitats.

In the sediment core from Gerzensee, Daphnia and Ceriodaphnia ephippia show very similar variations in their δ13C values in the part of the core where both groups are present (0–27 cm; Fig. 5). This suggests that Daphnia and Ceriodaphnia were subject to similar changes in their diet over time. In contrast, Plumatella statoblasts show a peak in δ13C values that does not coincide with the peak in the ephippia of Daphnia and Ceriodaphnia. The two groups also differ in the range of their δ13C values, which points towards different food sources for Plumatella and the cladoceran taxa. Access to different carbon sources may occur because of differences in both habitat and particle-size preference. Plumatella are sessile organisms in shallow areas of lakes, whereas free-living invertebrates like Daphnia and Ceriodaphnia can reach deep-water food sources. Zooplankton in deeper water layers may feed directly on MOB, or on other organisms that feed on MOB (e.g. ciliates), which can lead to distinctly negative zooplankton δ13C values (Jones and Grey 2011). In addition, Kaminski (1984) estimated that Plumatella can ingest particles between 5 and 17 μm in diameter, whereas Daphnia are able to filter particles between 0.5 and 30 μm in diameter (Geller and Müller 1981). Hence, bacteria (<2 μm) that are not attached to larger particles may be too small to be caught in the lophophores of Plumatella, but may be accessible as a food source for Daphnia. An alternative explanation for the observed differences between cladoceran and bryozoan remains may be that both groups rely on the same carbon sources (e.g. algae, detritus, heterotrophic bacteria, and possibly MOB), but their relative importance in the diet is different. Whether one, both or more reasons are causing the observed differences in δ13C values cannot be conclusively answered by this study.

Conclusions

In Gerzensee, Daphnia δ13C values closely follow the seasonal cycle of POM δ13C values, with the most negative values of Daphnia, down to −44.2‰, measured in early spring. Interestingly, Daphnia ephippia δ13C values did not show any pronounced seasonal variations. Seasonal changes in Daphnia δ13C values, and hence potential seasonal changes in the diet of Daphnia, are therefore not recorded in the δ13C values of Daphnia ephippia in the sediments of Gerzensee. Very low δ13C values of Ceriodaphnia ephippia in the flotsam, especially in winter (−50.1‰), confirm that there is a contribution of CH4-derived carbon to the diet of the crustacean zooplankton in Gerzensee. Plumatella statoblasts are less depleted in 13C than the two cladoceran taxa, indicating little or no influence of CH4-derived carbon in the diet of the bryozoan, at least during the time(s) when statoblasts are produced.

The average Ceriodaphnia and Daphnia ephippia δ13C values in the flotsam agree very well with the values in surface sediments, and for Ceriodaphnia, with measurements from a sediment trap. This shows that in Gerzensee the δ13C values of ephippia in the surface sediment, and potentially in general at a certain sediment depth, reflect an integrated value for ephippia floating on the lake during a particular time span, possibly over 1–2 years. In contrast, average flotsam Plumatella statoblast δ13C values were about 4‰ more negative than the surface sediment measurement. The reason for this mismatch may be an under-representation of the summer period in our data set. Nonetheless, this issue deserves further investigation in Gerzensee and other lakes.

Low δ13C values of the two cladocerans throughout the 62-cm-long sediment core, comparable to those measured in the flotsam, indicate that in Gerzensee, Daphnia and Ceriodaphnia relied on a 13C-depleted carbon source to supplement their diet throughout the past 150 years. Daphnia and Ceriodaphnia ephippia δ13C values showed similar variations, with the highest values for the two taxa around the same depth. This suggests that Daphnia and Ceriodaphnia have been subject to similar changes in their diet over time. In contrast, Plumatella statoblast δ13C values in the sediment record were again distinctly less negative, and showed a 3‰ increase at a different depth than the δ13C values of the two cladocerans. This difference might be explained, in part, by the feeding behaviour and different habitats of the organism groups, which influences the extent to which they can incorporate CH4-derived carbon. Our study confirms the findings of earlier studies, which indicated that taxon-specific δ13C values measured on aquatic invertebrate remains can provide insights into long-term changes in the relative importance of different carbon sources. This technique can be applied to better understand the impacts of, for example, land-use change, eutrophication, and climate change on the carbon cycling in lakes. Additional studies similar to the one presented here will be needed to further constrain the effects of seasonality on the δ13C values of fossil statoblast and ephippia samples, and to assess the extent to which seasonal patterns observed in Gerzensee are representative of those in other lake ecosystems.