Do subfossil Cladocera and chydorid ephippia disentangle Holocene climate trends?

Study site

The study lake, Lake Arapisto, is a small (~ 2 ha, max. depth ~8 m) dystrophic lake, which is situated about 70 km northwest of Helsinki (60°35′N, 24°05′E) at an altitude of 133 m a.s.l. (Figure 1). Basic limnological properties (dissolved oxygen, specific conductivity, pH, color, total nitrogen, and total phosphorus) are shown in Table 1. The lake is located in the southern boreal vegetation zone, dominated by Picea abies, Betula spp. and Alnus glutinosa. The catchment of the lake is characterized by bedrock outcrops (metamorphic rocks), tilly soil and paludified depressions. The aquatic vegetation consists of Phragmites australis, Nuphar lutea and Menyanthes trifoliata growing in the littoral zone. The basin consists of two steeply deepening depressions.

OPEN IN VIEWER

Figure 1.

Location of the study sites across Finland together with geographical settings and bathymetry of Lake Arapisto in southern Finland

OPEN IN VIEWER

Table 1.

Limnological parameters from Lake Arapisto measured in winter 1994 at depths of 1 and 7 m (HERTTA database, Finnish Environment Institute)

	1 m	7 m
Dissolved oxygen (mg/l)	5.1	0
Specific conductivity (µS/cm)	41	42
pH	5.5	5.6
Color (Pt mg/l)	100	140
Total nitrogen (µg/l)	630	920
Total phosphorus (µg/l)	22	53

The Arapisto basin is located northwest (~5 km) of the marginal moraine Salpausselkä II (SSII), which was formed by the oscillation of Scandinavian Ice Sheet (SIS) during the Younger Dryas (YD) at 12 850–11 650 cal. BP (Rasmussen et al., 2006). During the YD the Arapisto basin was under the SIS. At the onset of Holocene, around 11 590 cal. BP, SIS retreated some kilometers northwest from the SSII but still covered the Arapisto basin (Palmu, 1999; Tikkanen and Oksanen, 2002). Under the warming climate of the early Holocene and under the deglaciation, SIS started to retreat towards the northwest and Lake Arapisto was formed on a supra-aquatic ground (not influenced by the past developmental freshwater and marine stages of the Baltic Sea) ~11 500 cal. BP in harsh periglacial environment.

Lake Hirvijärvi, which is discussed in the current study, is located ~80 km east of Lake Arapisto (Figure 1). The geographical settings of Lake Hirvijärvi, where from the chironomid record originates, are given in Luoto et al. (2010).

Sediment sampling and Cladocera analysis

The 267 cm long sediment sequence from Lake Arapisto was cored from the eastern part of the basin (Figure 1) with a Livingstone piston corer through the ice in March 2000. The water depth at the sampling point was 6.9 m. The obtained sediment core consisted of silt (267–263 cm), silty gyttja (263–230 cm), and fine detritus gyttja (230–0 cm). The sediment core was sliced in the laboratory at 5 cm intervals until 260 cm and the last samples were sliced at 2–2.5 cm resolution. The age estimates of the sequence are based on seven radiocarbon dates and the age–depth model is published originally by Sarmaja-Korjonen and Seppä (2007).

For the subfossil Cladocera analysis, 1 cm³ of wet sediment was heated and stirred for approximately 20 min in 10% KOH and washed through a 44 µm mesh (Korhola and Rautio, 2001; Szeroczyńska and Sarmaja-Korjonen, 2007). The residue was concentrated by centrifuging and the samples were mounted in glycerine jelly on microscope slides. A minimum of 200 identifiable exoskeletal remains (carapaces, headshields, post-abdomens, ephippia) were counted from each sample and identified following descriptions by Szeroczyńska and Sarmaja-Korjonen (2007). The most abundant body part was chosen to represent the number of individuals for each species and the relative proportions for all taxa were calculated from this sum of individuals. During the Cladocera analysis subfossil mandibles of Chaoborus (phantom midge) were also counted from the same microscope slides to evaluate its presence in the lake. The abundance of Chaoborus was calculated from the total sum of Cladocera (Sarmaja-Korjonen, 2002).

For ephippium analysis (Sarmaja-Korjonen, 2003, 2004) chydorid carapaces (representing asexual reproduction) and ephippia (representing sexual reproduction) were enumerated from the same samples during the routine counting of Cladocera. Ephippia of planktonic taxa within genera Daphnia and Bosmina also preserve in lake sediments but they cannot be used in estimating proportions of asexual and sexual reproduction because their asexual remains (carapaces) do not preserve (Daphnia) or are not always identifiable to species level (Bosmina). The relative proportions of ephippia of individual species and total chydorid ephippia (TCE) were calculated from the number of chydorid carapaces and chydorid ephippia to represent the proportion of sexual reproduction of all chydorid reproduction. Ephippia of Alona affinis and Alona quadrangularis look very similar and it was not possible to reliably identify them separately, and therefore they were individualized as A. affinis/quadrangularis ephippia.

Data analyses

All data analyses were performed using relative cladoceran abundances. Zonations of the Cladocera assemblages and chydorid reproduction in the core from Lake Arapisto were performed with cluster analysis, using the program PAST (Hammer et al., 2001). The cluster analysis was run with constrained unweighted pair-group averages, where clusters are joined based on the average distance between all members in the groups. The cluster analysis was performed using the Euclidean distance as a dissimilarity measure, and the separate faunal zones were considered to have distances >25.

A Cladocera-based Finnish mean July air temperature transfer function (Luoto et al., 2011) was applied on the Cladocera assemblages to create a quantitative estimate on past July air temperature (T_JulyCla) variability during the Holocene. The T_JulyCla inference model was developed from a 76 lake data set across Finland, with the WA-PLS technique using two regression calibration components. The training set lakes are all small (0.003–1.419 km²) and shallow (sampling depth 0.5–7.0 m) and were chosen for the study to represent different climatic conditions. The range in mean July air temperature is 11.3–17.05°C and in mean annual air temperature −2.04 to 5.81°C. Altitudinal range is from 11.3 to 404 m a.s.l. The model had coefficient of determination (r²_jack) of 0.67, root mean squared error of prediction (RMSEP) of 0.86°C and mean and maximum biases −0.018°C and 1.733°C, respectively. The reconstruction was performed using the program C2 (Juggins, 2007). Canonical correspondence analysis (CCA) was used to assess the applicability of the training data for the core samples from Lake Arapisto for the evaluation of the reliability of the reconstruction (Bigler et al., 2003). The CCA was run with square-root transformation of the species data and down-weighting of rare species and performed with the program CANOCO, version 4.52 (ter Braak and Šmilauer, 2002).

Cladocera and chydorid ephippia

A total of 35 cladoceran taxa were identified from the sediment core of Lake Arapisto, the most abundant taxa (maximum occurrence > 3%) are illustrated in Figure 2. The Cladocera stratigraphy was divided into five local faunal zones (Cla1–Cla5) according to the cluster analysis (Figure 2). The cladoceran fauna in zone Cla1 (~11 000–10 500 cal. BP) included planktonic Bosmina (Eubosmina), pioneer chydorids Alona quadrangularis, Chydorus piger, Alona intermedia, and Rhynchotalona falcata. Zone Cla2 (~10 500–7500 cal. BP) was characterized by high abundance of Eubosmina and stable occurrence of chydorids A. nana, A. excisa, and A. affinis. Alona quadrangularis, A. intermedia, and C. piger decreased ~9500 cal. BP onwards. Acroperus harpae started occurring continuously from ~10 000 cal. BP onwards and Bosmina longirostris from ~9500 cal. BP onwards. In zone Cla3 (~7500–5000 cal. BP) B. longirostris increased markedly and proportion of Eubosmina diminished simultaneously. Littoral community remained stable. In the transition from zones Cla3 to Cla4 (~5000 cal. BP) planktonic community exhibited increase in Eubosmina and Daphnia. Simultanously B. longirostris decreased and disappeared from the community. Bosmina longirotris occurred in the zone Cla4 occasionally after its disappearance and started to occur continuously in the top of the zone. The topmost sample was distinctive representing the zone Cla5, as B. longirostris increased dramatically, Eubosmina decreased, and Daphnia disappeared. Dipterous aquatic larvae of Chaoborus occurred between 8000 and 2000 cal. BP in the samples.

OPEN IN VIEWER

Figure 2.

Relative abundances of Cladocera (maximum occurrence > 3%) and Chaoborus in Lake Arapisto, southern Finland

The number of ephippia encountered in the samples varied between 0 and 19 (mean 3) and the number of carapaces varied between 17 and 299 (mean 105). Ephippia of only 14 chydorid species were encountered, although carapaces of 24 species were identified from the sediment core of Lake Arapisto, of which the most abundant (ephippium proportion > 3%) are shown in Figure 3. The stratigraphy of chydorid carapaces and ephippia (indicating asexual and sexual reproductive modes) was divided into four local zones (Eph1–Eph4). The early-Holocene sediments until ~9500 cal. BP were characterized by high proportions of ephippia of Alona affinis/quadrangularis, Alona intermedia, and Chydorus piger (Figure 3). The TCE was very high in the lowermost sample with a maximum 26.1% at 267 cm and after the lowermost sample the TCE decreased gradually to c. 16% (at 265 and 262.5 cm), indicating the transition from zone Eph1 to Eph2. Zone Eph2 was characterized by variable TCE proportions of around 6–10% (~10 000 cal. BP) and the zone Eph3 TCE proportions around 5%. Ephippia occurred sporadically in the zone Eph4 and were completely absent in some samples during ~7500 and ~5000 cal. BP. At ~3000–2000 cal. BP ephippia of Alonella excisa and Graptoleberis testudinaria increased, elevating the TCE slightly.

OPEN IN VIEWER

Figure 3.

Relative abundances of chydorid carapaces (asexual reproduction) and ephippia (sexual reproduction) of the most common species with ephippium production (maximum occurrence of ephippia > 3%) in Lake Arapisto, southern Finland. Note the different scales between carapaces and ephippia

T_JulyCla reconstruction

The CCA ordination for the surface sediment and core samples (Figure 4) showed that the passively plotted core samples were enveloped inside the range of the training set, suggesting that the training set is valid for temperature inferences from the Lake Arapisto core. The T_JulyCla reconstruction showed low temperatures (~13.5–12.5°C) during the initial stage of Lake Arapisto ~11 000–10 500 cal. BP (Figure 5). The inferred temperature curve rose to 15°C until ~9500 cal. BP and later between ~6000 and 4000 cal. BP to 16°C. The period of 3000–2000 cal. BP was characterized by slightly (~0.5°C) lower temperature inferences. The inferred temperature value of 16.3°C for the topmost sample is identical to the modern measured mean July air temperature. Figure 5 also illustrates the pollen-based T_Ann reconstruction from Lake Arapisto from the same sample depths (5 cm resolution) as Cladocera analysis was performed. This reconstruction, with high resolution data at depths 186.5–163 cm was originally published by Sarmaja-Korjonen and Seppä (2007). The curve for total chydorid ephippia is also plotted in Figure 4, showing consistent trends with the quantitative temperature records. Another comparative record is that from Lake Hirvijärvi (Luoto et al., 2010), which also shows similar trends than reconstructed from Lake Arapisto (Figure 4).

OPEN IN VIEWER

Figure 4.

CCA ordination of the training set Cladocera assemblages (white circles) across Finland and core assemblages from Lake Arapisto (black circles)

OPEN IN VIEWER

Figure 5.

Holocene temperature reconstructions and the TCE curve from lakes Arapisto and Hirvijärvi, southern Finland. A locally weighted scatterplot smooth with a span 0.2 is applied on the records (grey line). The arrows in the top indicate present-day temperatures and error bars indicate the models’ prediction error for temperature reconstructions

Faunal development

The pioneer and early Holocene Cladocera fauna in Lake Arapisto until ~9500 cal. BP (Figure 2) consisted mostly of chydorid species Chydorus piger, Alona intermedia, Rhyncotalona falcata and Alona quadrangularis, which live on minerogenic bottom substrata and are typical postglacial pioneers in Finland (Sarmaja-Korjonen et al., 2003, 2006). The development of littoral vegetation most likely proceeded from ~9500 cal. BP, when the annual and summer temperature had risen to values comparable with today (Figure 5), allowing aquatic vegetation to flourish. This is also indicated by the appearance of vegetation-associated taxa, such as Graptoleberis testudinaria and decrease of the pioneering benthic species (Figure 2).

Major changes occurred in the plankton community ~7500–5000 cal. BP when Bosmina longirostris replaced Eubosmina as the dominant taxon at the transition from zone Cla2 to Cla3 (Figure 2). Simultaneously, remains of predatory midge larvae Chaoborus appeared into the stratigraphy and increased markedly during the mid Holocene (Figure 2). Since occurrence of Chaoborus in small lakes is usually without fish in southern Finland (Luoto and Nevalainen, 2009), these faunal changes may have been driven by changes in local fish populations. Furthermore, Chaoborus is known to prey on adult and larger sized cladocerans (Mumm, 1997) and therefore it may have favored preying on Eubosmina and the smaller bosminid B. longirostris would have succeeded because of its smaller size (Brooks and Dodson, 1965).

Late-Holocene faunal development in Lake Arapisto was characterized by the disappearance of B. longirostris, and the succession and introduction of Eubosmina and Daphnia, respectively. The obvious change in planktonic cladocerans from small-bodied B. longirotris to large-bodied Eubosmina and Daphnia may have been driven by additional changes in the food-web structure as Chaoborus decreased simultaneously. With this respect, large-bodied taxa Eubosmina and Daphnia, being more efficient grazers, would have displaced the smaller planktonic species. The topmost sediment layer, being distict, showed high increase in B. longirostris proportions and disappearance of Daphnia and may reflect recent anthropogenic perturbations (Szeroczyńska, 1991, 1998).

Holocene temperature record

This study is the first attempt, known to the authors, to reconstruct Holocene temperature changes based on subfossil Cladocera assemblages (Figure 5) and showed roughly the major and consistent trends in Holocene temperatures of northern Europe which have been established numerous times earlier (reviewed in Birks and Seppä, 2010). The inferred T_JulyCla record from Lake Arapisto indicated cold climate (T_Jul ~12–13°C), but rapidly and steadily rising temperatures during the early Holocene; ~2°C between ~10 500 and ~9500 cal. BP. This T_JulCla record from the early Holocene is highly similar to a chironomid-inferred temperature record from an adjacent lake in southern Finland (Luoto et al., 2010; Figure 5), both suggesting similar climatic conditions for the early Holocene as today in northernmost Finland. Furthermore, the T_Ann record (Sarmaja-Korjonen and Seppä, 2007; Figure 5) from Arapisto, as well as another pollen record from south-central Finland (Heikkilä and Seppä, 2003) suggest extremely low (T_Ann ~−2 to 3°C) temperatures during the early Holocene and a rapid increase to temperatures similar to today (T_Ann ~ 4°C) until ~9500–8500 cal. BP. The harsh climate in south Finland during the onset of the Holocene is reasonable since the retreating SIS located near the basins (Tikkanen and Oksanen, 2002) and set up periglacial environmental conditions.

The stable and warm ‘optimal’ climate conditions during the Holocene thermal maximum (HTM) are evident from the previously published records from south Finland (Heikkilä and Seppä, 2003; Heikkilä et al., 2010; Luoto et al., 2010; Sarmaja-Korjonen and Seppä, 2007) and suggest warmer temperatures than present. However, although there were major faunal changes in Lake Arapisto during the HTM (Figure 2), the T_JulyCla record does not clearly indicate the established climate development but showed July temperatures similar to the present (Figure 5). Furthermore, the decreasing late-Holocene temperature trend from ~4000 cal. BP onwards (Heikkilä and Seppä, 2003; Luoto et al., 2010; Sarmaja-Korjonen and Seppä, 2007) was not represented in the T_JulyCla record.

The present T_JulyCla record (Figure 5) implies that cladocerans may not be the most sensitive paleoindicators for temperature inferences, because the T_JulyCla temperature curve was relatively monotonic after the early Holocene with no major increase during the HTM or decrease during the late Holocene. Rather, the current temperature inference suggests that cladoceran communities do not respond sensitively to air temperature and, maybe for this reason, have not thus far been used extensively as quantitative temperature indicators, whereas pollen and chironomids respond mainly and perhaps more directly to air temperature. The attention in Cladocera-based paleoclimatological studies has been given to short-term, major, and abrupt climate-change periods, such as the Lateglacial and early Holocene transition and the 8.2 ka event. Previously, Lotter et al. (2000) used the Swiss transfer function by Lotter et al. (1997) to infer summer (June–August) temperatures during the Lateglacial–early Holocene from chydorid assemblages in Lake Gerzensee (Switzerland). They compared their results with an oxygen isotope record and pollen-based temperature reconstruction and found similar trends. A duplicate cladoceran-based study of the same time period was conducted by Duigan and Birks (2000) in the Kråkenes basin (Norway) indicating rapid decrease in summer temperatures during the YD and a progressive increase in temperatures at the onset of the Holocene. Recently, Sarmaja-Korjonen and Seppä (2007) showed that high-resolution series of proportions of total chydorid ephippia (TCE) and a planktonic taxon Bosmina longirostris, which favors eutrophic sites, responded to rapid temperature decrease during the 8.2 ka event in Lake Arapisto because of decreased phytoplankton productivity and shorter open-water seasons. Consistently with these results, Szeroczyńska and Zawisza (2011) reported dramatic declines in B. longirostris proportions in three Polish basins and linked the declines to climate deterioration and lower fertility of the lakes during the 8.2 ka event. In addition, Kattel et al. (2008) inferred recent temperature changes based on subfossil Cladocera assemblages under the climate warming of the last decade from a remote Scottish mountain loch, although the relationship to instrumental climate observations was relatively rough. Furthermore, other environmental forcer, such as biotic interactions with fish and other predators within the food-web may be more pronounced in structuring Cladocera communities, as is evidenced by for example Davidson et al. (2007, 2010).

Chydorid reproduction during the Holocene

The present TCE results (Figures 3, 5) show high coherence to the established Holocene climate trends and the T_JulCla record (Figure 5), as the TCE and the temperature inferences were highly similar. The idea behind using TCE curve as a proxy for past changes in climate (temperature, open-water season) is that in northern or high-altitude lakes chydorids reproduce most of the open-water season asexually (indicated by carapaces in sediment layers) and an environmental stress, mainly the oncoming winter, triggers sexual reproduction, which produces resting eggs enclosed by ephippia (indicated by ephippia in sediments). Accordingly, the TCE curve is indicative of the relative proportions of asexual and sexual reproduction in chydorids during the open-water season, which are linked to climate; during a mild climate the open-water season is longer and chydorids reproduce most of their active period asexually. In Lake Arapisto the high TCE values (~16–26%) of the lowermost samples during the early Holocene until 10 500 cal. BP (Figure 5) most likely indicate a very short open-water season in a periglacial environment near the retreating SIS margin, with a highly limited asexual reproduction period for chydorids. Similar TCE values occur currently in northernmost Finland (Sarmaja-Korjonen, 2007), indicating the important role of sexual reproduction and resting eggs under harsh climate. The gradually decreasing TCE until ~9500 cal. BP suggests that the asexual reproduction became slowly the overriding reproduction mode in chydorids after the cold early Holocene. Presently, chydorid sexual reproduction is limited to autumn months in southern Finland (Nevalainen, 2008) and this is shown as low TCE proportions in surface sediment (Sarmaja-Korjonen, 2007).

Although the TCE remained low (< 5%) after the early Holocene from ~9500 cal. BP onwards and there were periods during the mid Holocene when ephippia were absent in the sediments, the Holocene TCE curve shows corresponding trends with the inferred temperature records (Figure 5). The low and periodical occurrence of chydorid ephippia from ~9500 cal. BP onwards, suggesting a warm climate and long open-water season, is in accordance with other evidence of the onset and characters of the HTM in southern Finland (Heikkilä and Seppä, 2003; Heikkilä et al., 2010; Luoto et al., 2010; Sarmaja-Korjonen and Seppä, 2007). Hence, the absence of chydorid ephippia from sediment samples may therefore be even indicative of short periods with no ice cover on lakes and diminished role of chydorid sexual reproduction, as chydorids are able to overwinter as asexual individuals and reproduce asexually during particularly mild winters (Green, 1966; Koksvik, 1995). However, since the number of carapaces and ephippia found were at times low, the absence and low occurrence of ephippia during HTM remains but a discussion.

The current TCE was strongly similar with the T_JulCla and the other temperature records during the Holocene (Figure 5) implying a significant role of sexual reproduction in the temperature tolerance of the chydorid taxa. As illustrated in the development of cladoceran communities in Lake Arapisto (Figure 2), chydorid communities remained unexpectedly stable after the early Holocene (~9500 cal. BP onwards), despite the dramatic shifts in the planktonic communities. The climate of the HTM and late Holocene favored an asexual reproduction strategy, with occasional, most likely autumnal, sexual reproduction, indicated by the high abundance of asexual remains and low abundance and sporadic occurrence of ephippia (Figure 3). The reproduction strategy of cyclical parthenogenesis relies on asexual reproduction as a mechanism for vigorous population growth under favorable environmental conditions and sexual reproduction as a mechanism for creating new genotypes and providing diapause under unfavorable environmental conditions (Frey, 1982). Therefore, it could be argued that during the Holocene chydorids executed this protocol in order to maintain their populations in Lake Arapisto. Hence, chydorids would have reproduced predominantly sexually in order to renew the genotypes (high TCE) under environmentally unfavorable periods (early Holocene until ~9500 cal. BP) and some other occasions, and relied on the refreshed genotypes after the perturbations and used asexual reproduction mode to maintain these new and favorable genotypes. Thus, the favorable combination of asexual and sexual reproduction in cyclical parthenogens, such as cladocerans, may have proven to be a vigorous tool for adaptation to changing temperature.

To conclude, although the use of subfossil Cladocera assemblages and proportions of chydorid ephippia retain great potential for Holocene paleoclimatological investigations, their use in climate reconstruction may be hampered by environmental changes more independent of climate, such as food-web changes. Furthermore, these reconstructions may be vulnerable to adaptation abilities and large environmental tolerances partly through their reproduction strategy, which combines the ecological advantages of asexual and sexual reproduction.